Thermo-smoothing of Wood and Wood Materials

Proceedings of the 20th International Wood Machining Seminar June 7 – 10, 2011 Skellefteå Sweden Anders Grönlund

Luís Cristóvão

(Editors)

ISBN: 978-91-7439-264-7 Skellefteå 2011 www.ltu.se

Welcome to the 20th International Wood Machining Seminar Dear Colleagues and Friends On behalf of the Organizing Committee, I am pleased to welcome you to Skellefteå for the 20th International Wood Machining Seminar (IWMS 20). Luleå University of Technology (LTU) Division of Wood Technology is proud to host the seminar that started some 48 years ago. It is a credit to the founders and past seminar hosts that the seminar continues to attract a wide range of participants from many parts of the globe. The objective of IWMS-20 is to provide a forum for leading international researchers and practicing engineers to present and discuss recent advances in wood cutting tools, processes and machinery. Primary objectives are practical information exchange and technical interaction among wood machining professionals. I will particularly thank Professor Gary Schajer that has helped us a lot in planning and organising the seminar. Without Gary’s help and experienced advices we would not all be together at this time. I will also thank the entire staff at LTU/Skellefteå that has helped me with a lot of different practical issues. Your dedicated work has brought the seminar into reality. To our attendees and participants, welcome to the 20th IWMS. Enjoy the light summer nights and the beautiful nature in Northern Sweden. Meet colleagues and friends, make sure to make new ones, and give plenty of encouragement to those just starting out in this challenging field. Yours sincerely Anders Grönlund Chairman of 20th IWMS Organizing Committee

International Advisory Committee Prof. Gary Schajer, Chair, University of British Columbia, Vancouver, Canada Prof. Roland Fischer, Dresden, Germany Prof. Anders Grönlund, Luleå University of Technology, Skellefteå, Sweden Prof. Nobuaki Hattori, Tokyo University of Agriculture and Technology, Tokyo, Japan Prof. Shogo Okumura, Kyoto University, Kyoto, Japan Prof. Frieder Scholz, Hochschule Rosenheim, Rosenheim, Germany Dr. Ryszard Szymani, Wood Machining Institute, Berkeley, USA Prof. Chiaki Tanaka, Matsue, Japan Dr. John Taylor, FPInnovations, Vancouver, Canada Prof. Arto Usenius, VTT Techincal Research Centre of Finland, Espoo, Finland Dr. Darrell Wong, FPInnovations, Vancouver, Canada Prof. Zhijiang Zhou, Nanjiing Forestry University, China Editor’s Note Authors are responsible for the quality of their paper. Neither IUFRO nor the Luleå University of Technology shall be responsible for statements and opinions advanced in this publication.

CONTENTS 1. Key – note Address Future Processing of Wood Raw Material A. Usenius

1

2. Novel Processes Oral Presentations Thermo-Smoothing of Wood and Wood Materials - Recent Results I. Fuchs; T. Pflüger Effects of Wood Cutting with Extreme Inclined Edges R. Fischer; C. Gottlöber; M. Oertel; A. Wagenführ; W. Darmawan Investigation of Laser Beam Cutting of Lightweight Sandwich Panels H. Delenk; J. Herold; H-P. Linde; C. Gottlöber; A. Wagenführ

11 22 30

3. Saw-blade Characteristics Oral Presentations Nonlinear Vibration of Clamped Saws R. Khorasany; S. Hutton Practical Measurement of Circular Saw Vibration Mode Shapes G. Schajer; M. Ekevad; A. Grönlund Wood Chip Formation in Circular Saw Blades studied by High Speed Photography M. Ekevad; B. Marklund; P. Gren

39 47 55

4. Tool Material and Tool Wear Oral Presentations Effect of Cutting Tools with Advanced PVD Coating on Reduction of Power Consumption T. Minami; S. Nishio Carbide Tipping and Fatigue Strength of Band Saw Blades F. Scholz; U. Heisel; J. Tröger; M. Großmann; A. Hemer; R. Beier; A. Sidharta; K. Cheng Optimization of the Choice of the Cutting Materials for Solid Wood Tools P-J. Meausonne; S. Auchet Wear Characteristics of Newly K10 Coated Cutting Tools in Cutting Particleboards W. Darmawan; D. Nandika; H. Usuki; R. Marchal Poster Presentations Measurement of the Cutting Tool Edge Recession with Optical Methods J. Sandak; P. Bartosz; K. Grzegorz Wear of Teeth of Circular Saw Blades M. Ekevad; B. Marklund, L. Cristóvão Agricultural Residues in Panel Production – Impact of Ash and Silica Content C. Müller; R. Deetz; U. Schwarz; V. Thole

63

71

80 88

97 107 114

5. Process Monitoring Oral Presentations Cant Shape Measuring System and Knives Wear Tests B. Laganière i

123

Detecting Top Rupture in Pinus sylvestris Sawlogs J. Skog; N. Lundgren; J. Oja Sensory Quality Assessment of Surfaces, Especially Wooden Surfaces A. Riegel; K. Dekomien; J. Baade Monitoring of Measurement Accuracy - an Industrial Case D. Englund Poster Presentations Development of Instant Identification Method of CCA-Treated Wood using LaserInduced Breakdown Spectroscopy Y. Aono; K. Ando; N. Hattori Detecting Board Defects with a Time of Flight Camera N. Björngrim; O. Hagman; E. Jonze FuLOG – Radio Based Data Logger for Integration in Production Processes K. Ahmad; M. Grotekemper; A. Riegel; S. Witte

132 141 149

155

160 166

6. Sanding – Surface evaluation Oral Presentations Modelling Wood Surface Geometry after Wood Machining J. Sandak Sanding with Coated Abrasives – Relations between Temperatures, Power and Surface Quality H-W. Hoffmeister; M. Luig; A. Fricke Surface Roughness of the Wood – Reality vs. Measured Figures E. Magoss; S.Tatai Poster Presentations Study on Distribution and Elimination of Static Charge on Laminate Flooring X. Ma; H. Zhou; T. Ding; X. Zhou; D. Shen Characteristics of Interior Décor of Hunting Lodge in Julin in Surface Quality Aspects A. Rozanska; I. Swaczyna Influence of Climate on Surface Quality of Antique Wooden Flooring in Manor House A. Rozanska; A. Tomusiak; P. Beer

173 181

189

197 202

208

7. The Sawing Process Oral Presentations Economical Wood Sawing With Circular Saw Blades R. Wasielewski; K. Orlowski; S. Szyszkowski Improvement of Sawing Efficiency in Sawing Frozen Wood Y. Ikami; K. Murata Influence of Sawing Pattern on Lumber Quality in Sugi Large Logs Y. Matsumura; K. Murata; Y. Ikami Increased Sawing Yield by Thinner Saw Blades and Adapted Green Target Sizes J. Flodin; A. Grönlund Poster Presentations Sawing Patterns for Sugi Large Logs Using at Sawmills in Japan K. Murata; Y. Ikami; Y. Matsumura

ii

219 227 234 241

247

Wood Processing Capacity of Sawmills and Carpentry Workshops in Ghana E. Appiah-Kubi; M. Adom-Asamoah; K. Frimpong-Mensah; S. Tekpetey Development of Sustainable Bamboo Industries in Ghana: The Factors that will Interplay S. Tekpetey; E. Appiah- Kubi Energy Consumption Structure of the Hungarian Wood Industry M. Varga; G. Németh; Z. Kocsis; I. Bakki-Nagy A Novel Approach for Log Sawing Optimization P. Cao; Y. Wang; F. Huang; X. Guo

255 262

267 275

8. Chip Formation Oral Presentations Application of Fracture Mechanics for Energetic Effects Predictions While Wood Sawing K. Orlowski; T. Ochrymiuk; A. Atkins The Relationship between Macroscopic Chip Type and Microscopic Crack Behaviour in Wood Cutting M. Koshizuka; T. Ohtani; M. Inoue Orthogonal Cutting as a Method for the Determination of Fracture Properties of Oriented Wood Tissue M. Merhar; D. Bučar; B. Gospodarič; B. Bučar Machining of Wood using a Rip Tooth: Effects of Work-piece Variations on Cutting Mechanics A. Naylor; P. Hackney; E. Clahr Poster Presentation Some Investigations on the Counter and Down Milling E. Csanády; S. Németh; I. Bakki-Nagy; E. Magoss

281

289

296

305

315

9. Component Production Oral Presentations Wood Material Features and Technical Defects that Affect the Yield in a Finger Joint Production Process O. Broman; M. Fredriksson Added Value and Technical Performance of Manufacturing Window Frame Components from Pruned Fast-grown Scots Pine Logs Using Combined Live Sawing and Parallel-to-Surface Sawing Technique E. Verkasalo; E. Sirparanta; S. Silvennoinen A Simulation Tool for the Finger-Jointing of Boards M. Fredriksson High Capacity, High Yield Dual Order Primary Breakdown P. Gjerdrum Poster Presentations Bondability Study of Three Guianese Woods for Glulam Manufacturing D. Bourreau; J. Beauchêne ; Y. Aimene ; O. Nait-Rabah ; B. Thibaut Traceability and Analysis of Production in the Wood Industry N. Hassen; A. Zerizer

iii

323

334

342 353

361 370

Evaluation of the Passive Impregnation Method Compared to Full Cell Method in Terms of Leaching of Copper Azole from Treated Lumber K. Ando; Md N. Islam; H. Yamauchi; N. Hattori Predicting the Propagation of Diving Grain in the Vicinity of Sound Knots M. Öhman; M. Chernykh

378

386

10. Sound and Dust Oral Presentations Sound Source Localization and Noise Control in Wood Machining H-W. Hoffmeister; E. Topbas Influence of Wood Heat Treatment on Cutting Tool Temperature, Surface Quality and Dust Size Distribution of Mongolian Pine Wood T. Ding; L. Gu; N. Zhu; X. Guo; W. He Basic Cutting Examination of B-PSL Made from Bamboo S. Nishio; Y. Nishikawa

397 405

413

11. Cutting Forces Oral Presentations Optimizing the Cutting of Tension Wood in Rubberwood: An Economic and Quality Perspective J. Ratnasingam; F. Scholz. Main Cutting Force Models for Two Species of Tropical Wood L. Cristóvão; O. Broman; A. Grönlund; M. Ekevad; R. Sitoe Lateral Cutting Forces for Different Tooth Geometries and Cutting Directions M. Ekevad; B. Marklund Cutting Forces for Wood of Lesser Used Species from Mozambique I. Lhate; L. Cristóvão; M. Ekevad; R. Sitoe Poster Presentations Effect of Crushing Condition on Crushing Energy of Lumber F. Kiyohiko; Y. Yuji; T. Takahiro; T. Tsutomu Model for Forces on a Cutting Tooth of a Circular Saw Blade for Wood Rip Sawing M. Ekevad; A. Axelsson; L. Cristóvão

419

428 436 444

453 459

12. Milling and Planing Oral Presentations High Speed Wood Machining of Spinning Friction in Condition of Tribology T. Ohtani; R. Iida; T. Nakai; K. Adachi; M. Inoue Fundamental Investigations for Optimizing the Geometry of Face Milling Cutters U. Heisel; S. Ivanova The Effect of Planing on Shape Deformations in Pine A. Axelsson Investigation on Kick Back Behaviour in Peripheral Milling U. Heisel; M. Schneider

iv

467 473 485 493

13. Joinery and Furniture Production Oral Presentations Machining Properties of Multi-Layer Window Scantlings F. Scholz; M. Keller; A. Fathollazadeh Quality Control and Grading of Sawn Timber from Mills in Southern Ghana: Constraints and Opportunities S. Tekpetey; M. Ofosu; E. Appiah-Kubi A Product Data Model and Computer Aided Manufacturing for the Process Chain of Profiling C. Kortüm; A. Riegel Influence of Veneer Lathe Checks on Flexural Properties of Laminated Veneer Lumber L. Denaud; L. Bleron; M. Krebs; R. Marchal Poster Presentations How to Consider about Swelling and Shrinkage Explicitly in Case of Tolerance Analysis within Wood Working K. Solbrig; A. Riegel Asymmetrical Veneering with Hardwood Species of Different Shrink Value S. Oleńska; P. Tarcicki; A. Cichy; M. Mamiński; P. Beer

v

505 514

520

528

535

543

1. Key-note Address Oral Presentations

20th International Wood Machining Seminar

Future Processing of Wood Raw Material Usenius, Arto1 Heikkilä, Antti1 Usenius Timo1 1

VTT Technical Research Centre of Finland, FINLAND ABSTRACT

There are many options to improve wood processing performance. Potential degree of improvements depends on many factors like quality of wood raw material and product specifications. Some aspects are presented in the following: Improvement of communication between building and interior designers and companies offers a channel to increase wood usage. This necessitates development of Electronic libraries of wood products have to support planning processes. The designers, customers and owners have to have direct communication and feedback channels with the woodworking companies. Business concepts based on networking taking good care about customers’, further conversions’ and end users’ needs. This approach enables to react quickly on changing circumstances. Customers’, end users’ and refineries’ service through developing information component as a part of products. Managing and optimisation material and information flows in overall ICT systems and production planning systems. Supporting information flows between different actors in wood chains. Optimisation of wood raw material use: wood raw materials, products and semi-finished with specific properties. Strong integration of technological, information and service platforms. Intelligent, flexible and self-learning scanning, production and logistic systems support realisation business. Generation of feedback information to be used in self-learning production systems. Developing product specifications and families (strength, appearance, aesthetic aspects) matching wood raw materials Early stage information will be available before buying wood raw material and before harvesting and crosscutting of stems. Stem terminal cross cutting. VTT developed WoodCIM® and InnoSIM model and simulation systems for optimization throughout the conversion chain. Some results presented in this paper clearly show that it is possible to increase sales value and profit of the conversion considerably by implementing new technologies and concepts in the production.

1


1. INTRODUCTION Typical production and business features of sawmilling companies are the following: Conversion from the forest to the customer is not an unbroken and smooth production chain. Delivery and processing time may require weeks or months. Volume output and cost minimisation are emphasized in the production. Production is not flexible allowing only marginal freedom. Production and business are not adaptive. Feedback information is not generated and thus cannot be utilised. Limited volume of reliable and less reliable data is measured, however only locally used. Product properties vary considerably due to the non-homogeneous wood raw material. It is not possible to produce only high-value products with desired and specified properties. Secondary “falling” products which are not desired, are inevitably produced along the manufacturing processes. Mismatch between wood products and available wood raw material recourses may cause big economic losses. Wood working industry is far behind from the other industries providing services. 2. ADAPTIVE OPTIMISATION OF ACTIVITIES THROUGHOUT CONVERSION CHAIN Traditionally different stages of the wood conversion chain have operated too much independently. In the conversion chain the product of the former phase provides raw material for the latter one. The raw material and semi-finished products are not optimal or even good in respect to the final product. The incompatibility between the wood raw material, conversion products and final product causes a lot of waste and considerable economic losses. The stages involved in converting the wood raw material into final products influence on each other as well as the result. To obtain a good economic result, the chain must be seen in its entirety. The wood raw material has to be chosen taking into account the requirements of the final products. This is the only way for optimal utilization of the wood raw material. The material flow proceeds from the forest to the customers. The information flows in the same direction but should also take the reverse course. Optimisation means determining the best possible solution within given constraints limiting freedom in business. In the wood products business there are always three main types of business constraints: Wood raw material, production capacities and wood products markets. There always have to be the criteria to be optimised – maximised or minimised. One of the most important criteria is annual profit. In the future criteria describing sustainable development have to be taking into account in the optimisation procedures. Concerning the wood chain, two types of optimisation can be identified: Global Wood Chain Optimization seeks to achieve maximum profit in the conversion chain – from the forest to the end product - in its entirety. In the global approach the supply chain phases are actively interacting ensuring the best possible economical result. Suboptimization means a procedure to achieve the maximum output in an individual phase of the conversion chain. In the suboptimization different phases in the chain are in a loose interaction or they have no interaction at all. From the economic point-of-view, global optimization is much more important than the suboptimization of individual process phases.

2


VTT Technical Research Centre of Finland has developed WoodCIM (Fig. 1) that is a model and software comprising seven modules for global optimisation purposes. The system describes the whole conversion chain from the forest to the end products. In industrial applications WoodCIM system is normally linked to administrative information systems at the sawmill from where input data is available. WoodCIM will deliver information for decision making purposes.

Figure 1: WoodCIM

consists of integrated software modules.

Important part of the future sawmill business is the creation of feedback information. In the WoodCIM procedure measured log or batch of logs characteristics are recorded to the system as an output data. This output is then compared with the estimated output through WoodCIM software. The procedure results to information of possible needs to change parameter values in the software. If yes,a new parameter values are determined and implemented in the software. This is the way to improve prediction accuracy. 3. INTELLIGENT MATERIAL FLOW CONTROL In individual phases of conversion, information is growing rapidly through measurements and observations (Fig.2). However this information is utilised only locally and will be lost afterwards. This happens all the way throughout the supply chain. It is not possible to link final products, raw materials and processing parameters together. Strong support for business development could be achieved if the lost information could be stored and utilised in the later conversion phases. Recovery of information can be achieved through marking pieces, reading of the markings and storing the data in a database for utilisation in different applications. Marking of pieces can be done using different techniques i.e. RF-tags and ink jet markings. Reading of the markings can be done by using antennas or cameras.

3


Marking - reading – information (MRI) system applications concern quality control, process control, planning procedures and customer service. Marking of pieces is also a way to show the origin of the pieces and can be for instance used to ensure that the material originates from a certified source. MRI provides a quite new approach for the management of material and information flows from forest to the end products supporting customer oriented business and added value production.

Figure 2: Recorded and lost information throughout the conversion chain. Some concepts are based on the collective following of wood material batches producing information on how to process certain categories or classes of wood. Other concepts are more detail oriented and necessitate following of individual pieces of wood raw material, semifinished products and final products

4. SCANNING AND MEASURING TECHNOLOGIES Future grading will be based more and more on customer specific requirements and hence also specific grading procedures will be needed. It is almost impossible for a human grader to adopt and change continuously new grading rules, which makes the customer specific visual grading particularly difficult. In order to avoid economical losses in standard grading and in order to support grading according to the individual needs of customers, we need machine vision based grading. In the grading supported by machine vision physical measurements are carried out concerning the surface properties or the internal characteristics of the wood material. Black and white, color or IR cameras are used in the measurements. The data is computerized into a map of defects, which in turn is input information for the software to determine the quality of the piece. X-ray scanning provides information for the internal characteristics of logs or sawn timber pieces. Log scanner systems for measuring shape and internal properties of logs can be used in the following processes: Log sorting station - optimisation of borders of log classes based on order files 4


Bucking and cross cutting terminal - optimisation of cross-cutting of stems and sorting of logs Just before sawing - optimisation of log rotation angle and sawing set up for individual logs The purpose of X-ray inspection system is to detect properties of round wood. The analysing software should be tailored to meet application requirements. Typical functions can include measurement of dimensions/volume, moisture content, volume of knots, rot and other defects and heartwood/sapwood ratio. In the future it’s possible to get a full description of wood properties i.e. individual annual ring structure and density profile. 5. SAWING METHODS AND POSITIONING OF LOGS Present production systems are effective, however bulk product oriented. They are not flexible and production of components with specified quality features and properties is difficult. Producing value added components – smaller pieces with specific dimensions and quality features – instead of standard products offers sawmills a big potential to improve profitability of sawn timber business. Production of components should be started directly from the logs. Three basic sawing methods, cant, live and profile sawing, are presented in Figure 3. Live sawing method can be the first phase in manufacturing of components. Component sawing method starts by live sawing operation (Figure 3). The flitches received in the first phase of sawing are transported to edging machines provided by scanning system and multi blade settings. Scanning will result to description of timber properties for optimisation value of wooden bars taking from the flitch. The bars are cross cut into desired lengths.

Figure 3: Principal presentation about cant, profiling and component sawing methods. Value yield as function of top diameter for cant sawing, live sawing and component sawing methods.

5


Three different sawing methods are compared in Figure 3. In the live sawing method the slice can be edged into one or two sawn timber pieces. In the component sawing method the slice can be divided into one or two bars, which can be crosscut into components. The best value yield is achieved by using the component sawing method 2, value yield is 170 €/m3 by top diameter of 250 mm. By using the live sawing method, value yield is 150 €/m3 when the top diameter is 250 mm and in the cant sawing method the corresponding value yield is 135 €/m3. Correct rotation angle of a log in sawing has a major impact on value yield of log sawing. A case study shows that the difference in value yield between the best rotation and the worst rotation is on average 15 percent for cant sawing and 10 percent for live sawing (Figure 4). The variations are due to the fact that different volumes and grades of sawn timber will be received on different rotation positions. Log shape and internal log characteristics have to be measured and taken into account in determining optimum log position in the sawing operations. The graphics in Figure 4 shows the impact of log rotation angle on value yield. Small difference in the angle may cause major change in yield. This is because of quality of sawn timber pieces which is determined by positions of defects i.e. knots in the piece. Typically difference between the highest and the lowest value is rather big.

Figure 4: Value yield as a function of log rotation angle for live sawing. Positioning of a log is not an easy task even for sophisticated devices. Always some rotation errors may occur due to mechanics. Loss in value yield as a function of rotation error for cant sawing method and live sawing method are presented in Figure 4. Functions are linear. An error of 10 degrees in standard deviation means 1,8% loss in value for cant sawing and 3,5% for live sawing. Those errors have to be taken into account in optimisation procedures. Positioning accuracy is also important in sawing of the cants (Figure 4). It is possible to achieve maximum yield if sawing operation can follow the form of cant. Straight sawing of curved cant may influence big economical losses up to 15 percent by sweep parameter of 20 mm. Positioning errors have to be taken into account in the optimisation procedures concerning all process control applications.

6


6. CONVERTING OF SAWN TIMBER INTO SPECIFIC COMPONENTS Standard sawn timber pieces can be upgraded through splitting them into smaller wooden pieces, components with specific quality features demanded by the customer. Figure 5 presents potential for value yield increase. The value increase is 150 % when lower C-grade sawn timber pieces are cut into components. When better A- and B-grade sawn timber pieces are upgraded into components, value yield increase is 40 %. In some pieces the value increase is negative which means that there is also a risk of losing money by upgrading. However this can be avoided by sorting sawn timber before processing.

Figure 5: Increase of value yield by upgrading bulk sawn timber into components through ripping and /or cross cutting. Finnish sawmill. 7. CROSS CUTTING OF STEMS AT TERMINAL Cross cutting of stems is very important part of wood raw material processing. The maximum value and volume yield is determined in bucking of stems. In later phases in conversion it is not possible to compensate the faults made in cross cutting of stems. Mismatch between available wood raw material and products to be manufactured may cause big economic losses. Modern harvesters are very sophisticated and effective. However they cannot measure accurately stem properties like internal knots. Best bucking and cross cutting result can be achieved when the stems or part of stems are transported to cross cutting station provided by x-ray scanner and advanced optimisation software. Stems or part of stems can, however, be transported to a terminal where bucking and crosscutting can be based on precise scanning of raw wood properties. This provides virtual, mathematical stems (Figure 6) that can be used in optimisation procedures capable to “fill the stems” with desired products. Logs are converted into products in primary and secondary conversion ensuring perfect match between raw materials and product. The terminal station can serve one mill or several mills. Several value chains can be identified between forest and end use of wood products. Phases in conversion chains are impacting each others. In order to achieve good results, the chain has to be considered in its entirety. There may also be interaction between parallel chains. In the optimisation of allocation of wood raw materials and wood flows through whole conversion system, all phases have to be taken into account simultaneously. Because of the profitability, the chain is depending on raw material, manufacturing systems and product specifications. That is why economics of a stem terminal is depending on phases up-stream and downstream. 7


Figure 6: The Optimising allocation of wood raw materials and wood flows. 1) Stands to be harvested, 2) Transporation of stems, 3) StemTerminal for optimised bucking and cross cutting of stems, 4) X-ray scanner at log sorting station, 5) Different primary sawing processes, 6 and 7) secondary conversion and 8) buildings. Table 1: Comparison between crosscutting operations with conventional harvesting machine and terminal crosscutting Performance Activity

Conventional harvester bucking and cross-cutting

Diameter

Mechanical – low accuracy

Bucking and cross cutting at terminal (StemTerminal concept) Accurate measurement

Log length

Mechanical – low accuracy

Accurate log length.

Detection of knots

Harvester operator’s visual estimation.

Surface and internal knots can be detected with good accuracy

Detection of heartwood

Impossible

High detection accuracy.

Other wood properties Optimisation

Impossible

Possible – depending on the scanning system configuration Optimisation is based on product demand. Customer specific product specification can be applied.

Deliveries of logs Impacts on processing capacities

Based on measured and estimated stem profile and on the demand and value matrixes for logs. It takes weeks to get ordered logs. Minimum time is few days. Difficult – almost impossible

8

Delivery of desired logs can be started and stopped just in time. Terminal can just in time to start of producing demanded log profile for maximising processing capacities.


Modern harvesters are very developed and sophisticated. However they are working in demanding circumstances in the forest. It is not possible to implement similar technologies i.e. scanning compared to terminal measurements. Comparison between crosscutting operations with conventional harvesting machine and terminal crosscutting results information presented in table 1. This table clearly shows that smart stem terminal cross cutting offers potential to improve profitability and customer orientation radically.

8. ADAPTIVE PROCESS CONTROL SYSTEMS In many woodworking processes there are several phases where material are handled in smaller pieces. The machines cut material e.g. by chipping producing desired products or semi finished products, waste, dust and noise. Energy is also required. Input data comprises machine, tool etc. related items. Output can be divided into two categories: products and their quality and human and environmental impacts. The adaptive control system comprise sensors for detecting dust and noise, quality of the products, sharpness of the knives, hardware and software for processing and storing the data, software based on cutting model for calculating optimum parameter values for cutting. The system will be provided with a subsystem capable to make the control system adaptive by storing cutting parameters and cutting results into the files and processing and converting the data into information and further to knowledge.

REFERENCES Usenius, A., Heikkilä, A., Song, T., Fröblom, J., and Usenius, T. 2010. Joustavat ja itseoppivat tuotantojärjestelmät sahateollisuudessa. VTT RESEARCH NOTES 2544. Espoo, Finland. ISSN 1455-0865. Available from http://www.vtt.fi/inf/pdf/tiedotteet/2010/T2544.pdf [accessed 20 April 2011]. Usenius, A., 2007. Flexible and Adaptive Production Systems for Manufacturing of Wooden Components. In Proceedings of 18th International Wood Machining Seminar. May 7-9, 2007 – Vancouver, Canada. pp. 187-196. Usenius A., Heikkilä A., 2007. WoodCIM® - model system for optimization activities throughout supply chain. COST E 44 Conference proceedings on Modelling the Wood Chain Forestry – Wood Industry – Wood Products Markets. September 17 – 19, 2007, Helsinki, Finland. pp. 173 - 183. Usenius, A., Niittylä K., 2002. Uusia mahdollisuuksia sahaustoiminnan ohjaukseen kappaleiden merkinnän avulla. (New possibilities for controlling activities in sawmilling business through marking of pieces). Puumies, 2002, No 9. FINNISH patent: Saukkonen E., Temmes J., Usenius A., Wahlström B., Method for control of piece flow in difficult production circumstances, 1983.

9


10

2. Novel Processes Oral Presentations


Thermo-smoothing of Wood and Wood Materials – recent Results Fuchs, Ingrid and Pflüger, Thorsten Institut für Holztechnologie Dresden, gemeinnützige GmbH (IHD), Zellescher Weg 24, 01217 Dresden, Germany [email protected]

ABSTRACT Thermosmoothing is a new finishing technology for wood materials, especially MDF. The furniture industry used the thermosmoothing with electrical heated tools. Producers of powdercoated and lacquered furniture have several advantages from the thermo smoothing procedure. Last time was tested the thermosmoothing of wood and there were some very good results. The thermo smoothed surfaces of different types of wood had a high smoothness and give a good basic of coating processes. The advantages of thermo smoothing by coating processes demonstrated extensive investigations. Thermo smoothed, sanded and brushed MDF surfaces were coated with powder. Various surface designs were chosen as model samples of real furniture elements. The effect of different preprocessing techniques were studied to obtain coatable surfaces. The model samples were produced in the IHD and the surface-properties of the model samples were tested there. The powder coating was carried out by an industrial equipment, which was suitable for Lowbake- and UV-powder. After its there were some tests to characterise the surface of the coated model samples, such as roughness, thickness of the layer, adhesion of the layer and chemical resistance. In addition there were some climate tests to investigate the danger to open cracks. It was found that the quality of pre-processed surfaces is a very important precondition on the final coating quality. Powder coating requires a flat surface with a homogeneous, high density. Therefore, special attention had to be paid to edges and inner profiles of furniture elements due to their lower density compared to the main surface. The results of the industrial powder coating experiments demonstrated the strong influence of the pre-processing technique onto the coating quality. It is showed that thermo-smoothing is the best pre-processing procedure for powder coating of MDF.

INTRODUCTION The precondition for a high quality of surface coating of wood and wood based materials, especially Medium Density Fibreboard (MDF) and wood is a very regular surface of the substrate. Regularity that means non only the smoothness, that means the density and the morphology. Wood and wood materials differ in their density; additionally, the morphological properties of wood are distributed non-homogenously. Traditionally, before coating the surface of wood and wood materials is sanded, which is affecting roughness, but not density and morphology. An alternative to the conventional finishing procedures is thermo-smoothing, a finishing technology for shaped surfaces of wood based materials developed by IHD [Rehm 2005, Fuchs 2007]. Firstly, a finishing technology using heat was applied to the treatment of veneer, using hot

11


press machine or irons. The mechanism of thermo-smoothing consists of heating of wood, which leads to plastification of lignin. If combining the heat treatment with pressure the plastified lignin is changing the morphological structure of the surface. The intensity of the change depends on time, temperature and pressure. The thermo-smoothed surface shows a uniform morphology in comparison with shaped and sanded surfaces.

1

PRINCIPLES OF THERMOSMOOTHING

Requirement of a good smoothing quality is a perfect conformity with the milled surface shape. Therefore the thermo-smoothing was integrated in the processing machinery (plant). Some of the main machines in the wood processing, especially in furniture production, are CNC-routers and edge processing centres. For both types of manufacturing lines smoothing tools and adapters were developed to be integrated into the plants. According to the different types of wood processing, there exist two types of thermo-smoothing tools and technologies:  Fix-smoothing for CNC-routers,  Roller-smoothing for edge processing. 1.1

Fix-Smoothing

The Fix-smoothing is adapted for CNC-routers. There are two principles of Fix-smoothing: - Spindle guided smoothing tools and - Smoothing with adapter. Booth variants work with an electrically heated tool. The temperature of the tool is controlled by a digital control unit. The implementation depends on the CNC-router and the production process. If the smoothing tool is guided by spindle the control of the temperature is integrated in the machine control unit. The process time is longer because more time is needed to change and to heat the tools. It is possible to work with a second spindle only for smoothing, but this is more expensive than to apply an adapter. The adapter has an additional tool mount and a separate digital control unit. The number of tools to be used depends on the number of profiles to be treated. Additionally, the number of profiles defines the number of tool changes. The tools can be changed manually or automatically by means of an automatic tool change system (Figure 1) integrated in the adapter. The surfaces are smoothed at a temperature between 300 °C and 400 °C. The feed rate of smoothing is comparable to the feed rate of the shaping process and varies between 1 m/min and 9 m/min. Temperature and feed rate depend on the specific material and the requirements on the quality of the smoothed surface. Figure 1 Smoothing adapter for automatic tool change (photo: INNOTECH) 1.2

Roller Smoothing

The roller-smoothing unit is suitable for edge processing centres. Due to higher feed rates of edge processing centres in comparison to CNC-routers roller-smoothing requires a greater contact

12


surface for heat transfer. Therefore the roller-smoothing consists of a series of smoothing tools. The smoothing tools are profiled rollers, which are electrically heated; the temperature is controlled by a digital control unit. Every roller owns one or more different profiles (normally three). The roller-smoothing unit is adjustable in height. Thereby it is possible to change the profile very fast. Edge-smoothing units are a speciality of roller smoothing. Angles of furniture doors often are not sharp, but the have a curvature with a small radius. The edge smoothing unit can be applied fort smoothing the curved angles. The smoothing temperature is similarly the temperature of Fix-smoothing and ranges between 300 °C and 400 °C. The feed rate ranges between 9 m/min. and 20 m/min. Higher velocities require additional smoothing tools. Figure 2 illustrates a roller-smoothing-unit with edgesmoothing unit for three profiles.

Figure 2 Roller-smoothing-unit (photo: AKE)

2

SURFACE-PROPERTIES OF THERMOSMOOTHED MATERIALS

Thermo-smoothing is carried out by a combination of smoothing, heating and compression of the surface layer of wood and wood based material. The affected surface layer can be characterised by a thickness between approximately 10 µm to 100 µm, which can be shown by scanning electron microscope (SEM). The finishing of surfaces is described by surface roughness. The roughness was determined by topographic parameters. Topographic parameters were obtained with an optical 3D-measuring devise of MikroCAD type GFM with a measuring area of size 12,5 mm x 9,5 mm and a vertical resolution of 1 µm. For a proper characterisation of surface roughness the parameters Sa, Sq, Sz, Sp, Sv, Sk, Spk and Svk were determined. The line roughness profiles were measured by an instrument of Perthometer type with a profile length of 56 mm. The measuring mechanical stylus has a radius of 5 µm.

13


2.1

MDF

Normally, the distinction between a milled and a smoothed profile can be visually identified. Figure 4 shows a sample of a milled and a smoothed profile. The surface roughness of this sample was measured by 3D-measuring device. The differing roughness of the profiles are shown in Figure 5 and 6. It is obvious, that smoothing reduces the peaks but not fills the pores.

Figure 4 Test sample with milled (right) and smoothed (left) profiles 35

250

MDF_milled

30

MDF_milled

MDF_smoothed

roughness [µm]

roughness [µm]

200 25 20 15 10

MDF_smoothe d

150 100 50

5 0

0 Sa

Sq

Sk

Spk

Svk

Sz

parameter

Sp

/Sv/

parameter

Figure 5 Roughness parameters of milled and smoothed MDF-surface

Figure 6 Roughness parameter Sz, Sp and /Sv/ of milled and smoothes MDF-surface

Various surface structures are demonstrated by the SEM in Figure 7. The effect of compressing is evident.

14


Figure 7 Surface of milled and smoothed MDF (right – milled, left – smoothed) (SEM, Bäucker, TU Dresden) 2.2

Wood Solid wood can be finished by thermo-smoothing, if the wood surface and the near-surface layer have no defects and irregularities. To demonstrate the opportunities of thermo-smoothing, tests were realised with different kinds of wood: beech, birch, oak and alder. Therefore, Fix-smoothing was used as appropriate finishing technology. Figure 8 shows a solid wood sample, which was milled and smoothed.. The difference of the roughness parameters obtained for milling and for Fixsmoothing is shown in figures 9 and 10.

Figure 8 Milled and smoothed solid wood sample 14

25

beech mi l l ed beech smoothed

12 beech milled

20

beech smoothed roughness [µm]

10

15

8

6 10

4 5

2

0 Ra

Rq

Rk

Rpk

Rvk

0 Rz

Rp

Rv

parameter p a r a me t e r

Figure 9 Roughness parameters of milled and smoothed beech surface

Figure 10 Roughness parameter Rz, Rp and /Rv/ of milled and smoothes beech surface Roughness in the curved part of the surface was obtained only by measuring two-dimensional roughness profiles. Details of the effect of thermo-smoothing of beech are demonstrated by SEM in Figure 11.

15


Figure 11 Surface of milled (left) and smoothed (right) beech (SEM) (SEM-picture, Bäucker, TU Dreden) Figure 12 demonstrates the compression effect of the thermo-smoothing. The thickness of the compressed layer is less than 100 µm. The intensity of the compression depends on the cell structure of the wood. The SEM and the roughness values demonstrate the effect of thermo-smoothing. The different structures of solid wood and MDF lead to different structures of thermo-smoothed surfaces (s. fig 7 and 11). The filamentous structure of MDF is highly modified by thermo-smoothing, whereas the native structure of solid wood is consistent. Nevertheless, for the coating of solid wood the thermo-smoothing shows the same effects as for MDF. Both for MDF and for solid wood the wood fibres do not respond when coated with liquid lacquers.

Figure 12 Details of a smoothes surface of beech

3

POWDER COATING OF DIFFERENT PREPAERED MDF

In the furniture industry MDF is used especially for furniture doors. Therefore, MDF is coated with liquid lacquers and in recent times with powder lacquer [Hauber 2003, Bauch 2005, Bauch 2007]. It is well-known, that the quality of the coating of even surfaces does not depend strongly on material properties like density. On the other hand the quality of coatings of curved surfaces, profiles and edges strongly depends on density of the surface layer, which often differs from the bulk density in the middle of the board. The density varies from surface layer to the middle layer of the board and is affected by milling and/or smoothing.

16


3.1

Pre-treatment of MDF

To investigate powder coating of furniture elements with complex surface shapes or holes a specially test sample was developed. It is shown in Figure 13. The length of the test sample is 500 mm and the width 350 mm. The curved edge has a radius of approximately 2 mm and the inner profiles have a specific contour, shown in Figure 14. The test samples were treated either by sanding or thermo-smoothing. The test samples were milled by a CNC-router of type MAKA SC 20N with 4 axes. The CNCrouter was used to smooth the inner profiles with a smoothing temperature of 320 °C and a feed speed of 3.5 m/min. The edge profiles were smoothed by a roller smoothing unit (produced by AKE) combined with HOMAG edge processing machine with a feed speed of the roller unit of 9 m/min and a smoothing temperature of 400 °C. The surfaces were sanded by a wide belt sanding (grinding) machine. The sanded edge profiles were processed manually. Figure 15 demonstrates the roughness of different variants whereas s means sanded (K240), fs fine sanded (K400) and ts - thermo-smoothed. Among the edge profiles the smoothed show the lowest roughness. The roughness of the smoothed inner profiles is comparable to that of fine sanded surfaces.

25 mm 22 mm 4 mm 6 mm

Figure 14: Ogee profile (Inner profile) Figure 13: Test component

17


roughness Rz [µm]

90

SF

80

IP

70

EP

60 50 40 30 20 10 E19_ts

E19_fs

E19_s

F19_ts

F19_fs

F19_s

B16_ts

B16_s

B16_6

A16_ts

A16_fs

A16_s

0

variant

Figure 15 Roughness Rz of the test components, SF = surface, IP = inner profile, EP = edge profile; A, B, E, F = different kind of MDF The powder coating experiments were carried out with UV-powder. Before coating, the fluidness, the bulk density and the flowability of the powder were determined. The flowability of the UVpowder was very low and, therefore, the application of the powder quite difficult. The hardening conditions of the UV-powder were defined before industrial experiments in laboratory tests [Mucha 2008]. The following tests with the test samples were carried out to characterise the effect of powder coating: -

3.2

subjectiv optical inspection, measurement of the thickness of the powder layer, measurement of the bond strengh with cross cut test (ISO 2409) on the edge profiles and the surfaces, measurement of the roughness of the inner profiles, and determination of the change of thickness and weight in the climate test ( 7 days 23 °C, 50 % rel. hum; 14 days 40 °C, 85 % rel. hum.). Roughness

The quality of powder coated surfaces mainly depends on the roughness and the density of the pre-processed surfaces. Figure 16 shows the roughness of coated and non-coated inner profiles for UV-powder-coating. The roughness of the profiles depends on the pre-processing and the density of the MDF-board.

18


80 Rz uncoated

roughness Rz [µm]

70

Rz coated

60 50 40 30 20 10 E19_ts

E19_fs

E19_s

F19_ts

F19_fs

F19_s

B16_ts

B16_s

B16_6

A16_ts

A16_fs

A16_s

0

variant

Figure 16: Roughness Rz of the non-coated and the UV-powder-coated inner profiles of the variants s, fs and ts

3.3 Climate Tests The climate test was carried out in two steps. At first a climate was defined with a temperature of 23 °C and relatively humidity of 50 % was set for a duration of 7 days followed by a climate of a temperature of 40 °C and relatively humidity of 85 % for a duration of 14 days. For the first step thickness and weight of every test sample were measured at the beginning and after 7 days. For the second step the measurements were done after 1, 4, 7, 11 and 14 days. The thickness was measured on 6 points of the samples, the results were averaged. The test results show, that on coated with UV-powder surfaces crack propagation occurred only in positions where they were initiated directly during or after coating or where no layer was faced. Samples without any cracks directly after coating do not show any crack initiation and propagation during climate tests. The changes of thickness and weight caused by alternating climates are demonstrated in Figures 17 and 18. It was observed that different MDF boards show different swelling behaviour. Figure 18 demonstrates that the swelling of the thermo-smoothed variants is lower than that of the others. But Figure 17 demonstrates that the influence of the MDF board and the powder coating is stronger than the processing [Fuchs 2010].

19


8

relative change of mass and thickness [%]

thickness 7

mass

6 5 4 3 2 1 0 A16_s

A16_fs

A16_ts

B16_s

B16_fs

B16_ts

variant

Figure 17 Relative change of thickness and mass of the MDF board A16 and B16 for all processing variants 6 relative change of thickness and mass [%]

s fs 5

ts

4

3

2

1

0 thickness

mass

Figure 18 Mean value of relative change of thickness and mass over all variants

20


4

SUMMARY

Thermo-smoothing of both MDF and solid wood leads to a distinct reduction of surface roughness. Due to the different material structure the roughness of MDF surfaces is reduced to a higher degree than the roughness of solid wood surfaces. In both cases the coating of the surface can be affected positively by thermo-smoothing. As a pre-treatment technology, thermo-smoothing improves the applicability of powder coating to MDF boards with UV powder with a high impact on surface quality and coating stability against climate changes (moisture and temperature). The most important challenge for powder coating is the quality of the MDF. This includes both the conductivity and the mechanical properties. If the powder lacquers have a low melting and hardening temperature with a short time required for melting, the danger of destroying the board is lower. Literature [Bauch 1999]

[Hauber 2003]

[Bauch 2005]

[Rehm 2005]

[Bauch 2007]

[Fuchs 2007]

[Mucha 2008]

[Fuchs 2010]

Bauch, H.: Elektrostatische Oberflächenbeschichtung von Holzwerkstoffen mit thermound strahlenhärtenden Pulverlacken. Dresden 1999, IHD-Eigenverlag, 78 S. Hauber, P.: UV-Pulverbeschichtung von MDF in höchster Qualität. JOT + Oberfläche, 43(2003)5, S. 12 – 17 Bauch, H.: Mechanische und thermische Vorbehandlung von MDF vor der Pulverlackierung; Vor- und Nachteile der verschiedenen Vorbehandlungsverfahren kennen. besser lackieren 7(2005)9, S.12 Rehm, K.; Raatz, C.: Developing of dust-free finishing processes for industry. Proceedings of the 17th International Wood Machining Seminar, Sept. 26 – 28, 2005, Rosenheim, Germany Bauch, H.; Emmler, R.; Krug, D.; Fuchs, I.: Powder coating of wood based materials – chanced, requirements on materials and application technologies, reachable surface qualities with UVhardening and thermosetting powder coatings. Drewno-wood 2007, vol 50, nr.177, p. 101 - 117 Fuchs, I.; Raatz, C.; Peter, M.; Pflüger, T.: Some special problems of thermo-smoothing and coating. Proceedings of the 18th International Wood Machining Seminar, May 7 – 9, 2007 – Vancouver, Canada Mucha, M.: Einflussfaktoren auf die Pulverbeschichtung von MDF, vornehmlich unter Verwendung von UV-Pulver. Diplomarbeit, 2008 Fuchs, I.; Kleber, D.; Mucha, M.; Pflüger, Th.: Einfluss der Vorbehandlung auf die Qualität beim Pulverbeschichten. Holztechnologie 51(2010), S. 34 - 38

21


Effects of Wood Cutting with Extreme Inclined Cutting Edges Fischer, Roland1; Gottlöber, Christian1; Oertel, Michael1; Wagenführ, André1 and Darmawan, Wayan2 1

2

Technische Universität Dresden, 01062 Dresden, GERMANY Agricultural University (IPB) Bogor, Bogor 16680, INDONESIA E-mail: [email protected]

ABSTRACT The inclination of cutting edges on cutting tools is a well-known method to improve the cutting behaviour when machining of wood and wood based materials. Since long a time this principle is applied on hand tools (planers) and on different machine tools for slicing, milling or drilling. The cause for are changes of the timing when the edge contacts the work piece. Also some changes regarding the edge angles in different planes are connected by the principle of edge inclination. Finally this influences the cutting power, the cutting forces and the tool wear. Well known is the potential to decrease the cutting noises when the edges are inclined. The paper tries to give some explanations about the mechanisms and the potentials when using tools with extreme inclined cutting edges based on theoretical observations and research results from the past. This article is a continuation of papers represented on IWMS 17 and 19. Keywords: tool inclination angle; cutting power; cutting force; tool wear; planing

INTRODUCTION The objective of cutting processes are useful products, high productivity, less energy efforts as well as less environmental impact. This goal is influenced by a multitude of factors which result a almost unmanageable field of configuration options and thus adjustment prospects. One of this options is given by the tool inclination angle λS. The angle is defined between the edge and the tool reference plane Pr measured within the tool edge plane PS according to the standards ISO 3002-1 or DIN 6581. The tool inclination angle λS is responsible for the so called drawing cut to get a smooth edge engagement during the cutting. For this reason the cutting force characteristic is changed and a decrease of the cutting noises can be obtained. Also some positive influences to the tool wear are known [1]. In the past at the Institut für Holz- und Papiertechnik of the Technische Universität Dresden the behaviour of procedures and tools using inclined edges was investigated on the example of linear cutting (slicing) [2] and milling (planing) [3]. Thereby the mentioned positive effects were confirmed. Based on former publications at the International Wood Machining Seminar [4, 5] and latest investigations general effects of wood cutting with extreme inclined cutting edges are introduced and founded in the following chapters. The main focus is on force, power and wear issues. The relations and dependencies of noise emission has been discussed in [5].

22


CHARACTERISTICS OF KINEMATICS AND GEOMETRY Procedures of rotary cutting (e. g. milling) and procedures of linear cutting (e. g. slicing) equal in terms of edge inclination, as a winding up of the edge on rotary tools leads again to a straight inclined edge. That is why in both cases the length of the active edge lS is only given by the inclination angle λs and the cutting width b (formula 1). The relation between the active edge length lS to the cutting width b can be calculated by formula 2 (figure 1, left).

b cos λS

lS 1 = b cos λS

(1, 2)

12

60

10

50

8

40

Velocity relation v a /v c

Length relation l s/b

lS =

6 4 2

30 20 10 0

0 0

10

20

30

40

50

60

70

80

0

90

10

20

30

40

50

60

70

80

90

Tool inclination angle λ S (°)


Figure 1: Relation between the active edge length lS and the cutting width b (left) resp. the engagement velocity va to the cutting velocity vc (right) depending on the inclination angle λS

On figure 1 (left) it is clearly to be seen that starting from an inclination angle of about 45° the active cutting length is raising disproportionately. If observing a cutting edge without inclination every point of the edge engages to the work piece during the same time (figure 2 a, left). That means the engaging velocity va of the cutting edge is nearly infinitely. An increasing of inclination brings a deceleration of edge engagement from point P1 to point P2 (figure 2 a, right). This delay can be expressed by the covered path lS,λ. In the corresponding time unit a assumed cutting edge of length lS,λ would effect with the engagement velocity va perpendicular to the direction of the cutting velocity vc.

Figure 2: Comparison of the engagement relations when cutting with different inclination angles λS on the example of linear cutting (a: geometry and kinematics; b: planes)

23


The relation between the engagement velocity va to the cutting velocity vc or the cutting width b to the length lS,λ is determined by the inclination angle λS according to formula 3. In this way the velocity va decreases with increase of inclination rapidly (figure 1, right). va b 1 = = vc l S ,λ tan λS

(3)

An important change of cutting behaviour is generated by the cutting edge angles when varying the inclination angle. Therefore three different tool planes are of importance (figure 2 b). If there is no inclination of the edge the angles within the tool orthogonal plane Po are equal to those of the cutting edge normal plane Pn. With increases of the inclination the wedge angles and the cutting angles within the tool orthogonal plane Po are going down. With constant normal cutting angle δn (formula 4) a raising inclination of the edge leads to a decrease of the orthogonal cutting angle δo (formulas 5, 6). The orthogonal wedge angle βo can be calculated based on the normal clearance angle αn, the normal wedge angle βn as well as on the inclination angle λS (formula 7).

δ n = αn + βn

⎛ tan(90° − δ n ) ⎞ ⎟⎟ ⎝ cos(λS ) ⎠

δ o = αo + βo

δ o = 90° − arctan⎜⎜

(4, 5)

(6) ⎛ tan(90° − α n − β n ) ⎞ ⎟⎟ cos λs ⎝ ⎠

β o = 90° − arctan(cos λs ⋅ tan α n ) − arctan⎜⎜

(7)

CUTTING POWER AND CUTTING FORCES Caused by the geometry and the kinematics of cutting with inclined tool edges an axial aligned force is generated increasingly. When cutting along the growing direction of wooden work pieces this force is oriented perpendicularly to the cutting direction. This part of the total cutting force F is called passive force Fp corresponding to the standards ISO 3002-4 or DIN 6584. That is why the emission of chips and particles of cutting procedures with rotary tools takes place in more axial oriented direction and with lower kinetic energy. To find out the conditions regarding the cutting power Pc and cutting force Fc when using tools with inclined edges basic investigations and experiments have been done [5, 6, 7]. With some tool prototypes the inclination angle λS was varied in a range between 45° to 85°. For comparison purposes also tools without edge inclination (λS = 0°) were prepared and used. The edge angles were constantly within the tool orthogonal plane Po. This leads to an increase of the normal cutting angle δn if the inclination angle raises. The results of the test runs showed a significant increase of cutting power Pc when 65° of edge inclination is exceeded. The power conditions when conventional milling of spruce (Picea abies) with constant cutting parameters are introduced on figure 3. One possible explanation of the behaviour can be given by the changed edge angles in the cutting edge normal plane Pn when the edge angles within the tool orthogonal plane Po are constantly.

24


That means the normal cutting angle δn increases when the inclination angle λS raises too. Normally that should lead to an increase of cutting power as measured. 500

100

Average engagement cutting force Fc,m

80

Example:

400

Cutting power Pc 60

300

40

200

20

100

Cutting power P c (W)

Cutting force parameter F c,total and F c,m (N)

Total average cutting force Fc,total

0

0

0

10

20

30

40

50

60

70

80

Conventional planing Work piece: spruce (Picea abies ) Cutting direction: B Edge material: Cold steel X155CrMo12-1 Edge hardness: 60 HRC Tool diameter d = 125 mm Number of edges z = 1 Cutting speed vc = 39.3 m/s Average chip thickness hm = 0.267 mm Feed per tooth f z = 1.733 mm Cutting width b = 30 mm

90


Figure 3: Cutting power Pc, total average cutting force Fc,total and average engagement cutting force Fc,m when planing of spruce (Picea abies) depending on the inclination angle λS

Another reason for the observed behaviour is the friction between the edge and the work piece material during cutting with inclined edges. With increase of the inclination a virtual movement along the cutting edge is imaginable. This assumed movement runs with the resulting sliding velocity vS of the vectorial interaction (formula 8) of the cutting velocity vc and the engagement velocity va introduced in the last chapter.

vS = vc + va 2

2

FS = ∆Fc + Fp 2

2

(8, 9)

The virtual sliding movement depends on the friction properties of the edge material as well as the work piece material and can be described by the dynamic friction coefficient µ. The resulting sliding force FS along the edge can be vectorial decomposed in two force components (formula 9) – the passive force Fp and a force ∆Fc (figure 4, left). This force ∆Fc effects against the cutting direction and can be added to the rest of the cutting force which is basically generated during the material cutting also without edge inclination. An increase of the edge inclination leads to an increase of the force component ∆Fc and that is why the cutting power Pc increases finally too. 12

Force relation ∆ F c/F p

10 8 6 4 2 0 0

10

20

30

40

50

60

70

80

90


Figure 4: Comparison of the engagement relations and forces when cutting with inclined edges on the example of linear cutting (left) and ratio between the forces when increasing the inclination (right)

Trigonometric considerations show that the relation between the passiv force Fp and the force ∆Fc is equivalent to the relation between the cutting width b and the virtual cutting length lS,λ or the corresponding velocities va and vc according to formula 10. With formula 4 follows the dependency to the inclination angle λS. Finally formula 11 shows the relation between the force

25


∆Fc and the passive force Fp depending on the inclination angle λS (figure 4, right). The function development seems to be similarly to the measurement results of the cutting power Pc (figure 3).

Fp ∆Fc

=

b lS , λ

=

va 1 = vc tan λS

∆Fc = Fp ⋅ tan λS

(10, 11)

Possibly in this way the cutting force Fc and the cutting power Pc increases because the force ∆Fc. The introduced theoretical assumptions should help to deliver an explanation for the obtained developments of cutting power when increasing the tool inclination angle (figure 4).

d

ϕλ ϕ e

lS

lS,λ

λS

lS,λ

ae

Sb b

circumference of the tool

0

2π feed f ϕλ+ϕe te te,λ

b

tm

λS

Fc

Fc,m Fc,total

Figure 5: Engagement angles, engagement lengths and duration of the cutting force Fc when milling with inclined cutting edges

The calculation of the cutting force parameter by using the measurement results of cutting power first requires an analytical consideration of the relations during the edge engagement and the effecting forces when the inclination of the edge raises (figure 5): When cutting with inclined edges the engagement time te is strongly increased by growing inclination angle λS. Therefore an extra engagement length lS,λ (formula 12) can be measured when the edge is cutting. This happens during the extra engagement time te,λ (formulas 13, 14). For rotary procedures an extra engagement angle ϕλ according to formula 15 can be declared. The standard engagement angle ϕe is given by formula 16 with respect to the cutting length Sb of one single edge dot (formula 17) and the tool diameter d. The cutting velocity vc can be calculated simply with formula 18.

lS ,λ = b ⋅ tan λS t e ,λ =

lS ,λ vc

(12)

te,λ =

26

b ⋅ tan λS vc

(13, 14)


ϕλ =

lS ,λ ⋅ 360° π ⋅d

ϕe =

⎛ 2 ⋅ ae ⎞ Sb = π ⋅ d ⋅ arccos⎜1 − ⎟ d ⎠ ⎝

Sb ⋅ 360° π ⋅d

vc = π ⋅ d ⋅ n

(15, 16)) (17, 18)

If the rotation number n and the cutting velocity vc is constantly then the measured cutting power Pc is a measure of cutting force Fc and the cutting energy Ec. The development of the cutting force depends on the engagement characteristic of the cutting edge. This can be characterized by three ranges. During the time of the first contact of the first edge dot with the work piece material until the completion of its cutting arc the cutting force raises to the maximum when cutting in conventional mode. After this the cutting force is nearly equal until the last active dot of the edge starts its cutting arc thru the work piece. Now the cutting force is going down quickly until the zero level is reached when the last active edge dot is out of the work piece. The cutting force which effects in the average during the edge engagement is defined as Fc,m. The assumed average cutting force of a total tool period (e. g. one tool rotation) can be defined as Fc,total (figure 5). The basis of the determination of the theoretical permanent average effecting cutting force Fc,total is the measurement result of the cutting power Pc with respect to the cutting velocity vc and the number of edges z (formula 19). The cutting energy Ec can be calculated with formula 20 by taking the cutting power Pc and the tool rotation number n into account. To calculate the average engagement cutting force Fc,m and the specific cutting force kc the following formulas 21 and 22 are used. Fc ,total =

kc =

Pc vc ⋅ z

Fc ,m b ⋅ hm

Ec =

Pc n

Fc ,m =

(19, 20)

Pc ⋅ 360° (ϕe + ϕλ ) ⋅ vc ⋅ z

(21, 22)

The load of the tool occurs with a very short impact and a high force when there is no edge inclination. This is because of all active edge dots approach to the work piece in the same time. The force Fc,m and the specific force kc becoming totally lower when the cutting tool is carried out with edge inclination but this force effects importantly longer (formula 22)! This effect happens because of a significant delay of the edge dot engagement when cutting. With the experimental results of the cutting power Pc (figure 3) and the formulas 19, 21 and 22 it is possible to calculate the values of Fc,total, Fc,m and kc. The obtained force developments depending on the tool inclination angle are also to be seen on figure 3. The progress of the total average cutting force Fc,total is like the progress of the cutting power Pc. This is not unexpected because Fc,total is proportional to the cutting power Pc, the cutting velocity vc and the number of cutting edges z (formula 19). This is true because the last both mentioned magnitudes have been constantly. Looking to the development of the calculated average engagement cutting force Fc,m or the specific cutting force kc the relations are totally different. With increases of the edge inclination these magnitudes decreasing strongly. That means the average load of the cutting edge, signed by the average engagement cutting force Fc,m, falls with increase of the edge inclination but effects

27


longer. This case is the explanations for the smoother cutting process of procedures with inclined cutting edges.

TOOL WEAR In experimental series [6, 7, 8] with some of the same tools as introduced for the cutting power measurements positive effects of edge inclination to the wear of cutting edges have been measured (figure 9). A increase of the inclination angle λS leads to a decrease of the cutting edge radius r. This effect was already found by Tröger [1].

Cutting edge radius r (µm)

50

Example:

40

Conventional planing Work piece: spruce (Picea abies ) Cutting direction: B Edge material: Cold steel X155CrMo12-1 Edge hardness: 60 HRC Tool diameter d = 125 mm Number of edges z = 1 Cutting speed vc = 39.3 m/s Average chip thickness hm = 0.154 mm Feed per tooth f z = 1.000 mm Cutting width b = 10 mm

30

20 lambda 0° lambda 65° 10

lambda 75° lambda 85°

0 0

200

400

600

800

1000

1200

1400

Cutting length lc (m)

Figure 6: Cutting edge radius r when planing of spruce (Picea abies) depending on the covered cutting path lc and for different inclination angle λS

With increasing inclination of the edges three changes are obtained which are responsible for the observed phenomenon: Firstly the active cutting length lS raises. With it the load during the cutting is distributed to a extended cutting edge. Secondly the smoothness of the cutting process increases as explained in the last chapter because of decreasing average engagement cutting forces Fc,m and in general lower maximum forces. One can say the load per length of the cutting edge decreases. Thirdly the edge inclination has consequences to the important edge angles, mainly in the tool orthogonal plane Po and in the cutting edge normal plane Pn (figure 2 b). A constant orthogonal wedge angle βo leads to a increased normal wedge angle βn with increasing inclination angle λS. This leads to a significant improved stabilization of the edge. The logical conclusion is a reduced tool wear as to be seen on figure 6. Another interesting aspect relevant to the tool wear is given by angle constellation on the edge. Because with the increase of edge inclination large normal wedge angles correspond to small orthogonal wedge angles. So brittle tool edge materials can be used for cutting applications which require rather small wedge angles in cutting direction (orthogonal plane). For example diamond as edge material can be used then for solid wood cutting when using large inclination angles. A very long cutting length should be possible before replacing the tool for maintenance.

CONCLUSIONS Cutting with inclined cutting edges is one possibility to influence the process behaviour mainly in terms of noise emissions, edge load characteristics, tool wear as well as chip collection positively. Until today this possibilities are used only in special applications. This is because of the complicated shape and construction of such tools when thinking about rotating tools. With

28


increasing of the tool inclination angle the manufacturing efforts raises. But the knowledge about the relations and the advantages should lead to more applications especially in planing of wood. A higher price of such tools can be balanced because the tool its improved behaviour. The main reasons for the improved properties of the cutting process with inclined cutting edges are the harmonic, smooth and time-delayed edge engagement of following edge dots along the edge. Furthermore a extended length of the active edge and important changes regarding the angles at the edge in different planes are responsible for changes. Regardless an increase of the cutting power when cutting with extreme inclined edges the average cutting force during the elongated edge engagement is going down to very low values. That is why the specific cutting force is also decreased. The kinematical changes of edge engagement leads to a fundamental changing in calculating the average engagement cutting force Fc,m and can be calculated with the introduced equations. With the principle of edge inclination the tool wear can be reduced significantly. The reduction of the average engagement cutting force leads to a reduction of the average load on the cutting edge too when increasing the tool inclination angle. Also a change of the wedge angles is responsible for improvements of wear resistance. In this case diamond as a cutting edge material which requires a large normal wedge angle can be used for solid wood cutting.

REFERENCES 1. Tröger, J. (1990) Einfluß des Neigungswinkels beim Umfangs- und Stirnfräsen. HOB Die Holzbearbeitung, 37 (6): 34-43. 2. Fischer, R.; Gottlöber, C. (1998) Verbesserung der Oberflächenbearbeitung von Holz mit geometrisch bestimmten Schneiden durch lineare Schnittbewegung. Final report of research project AiF-Nr. 10811 B/1, TU Dresden. 3. Wagenführ, A.; Gottlöber, C.; Oertel, M. (2010) Entwicklung von Schraubfräswerkzeugen für die Holzbearbeitung. Final report of research project AiF-Nr. 15424 BR, TU Dresden. 4. Fischer, R.; Gottlöber, C.; Rehm, K.; Rehm, C. (2005) A Milling Cutter As A Screw – Cutting Instead Of Hacking. Proceedings of the 17th International Wood Machining Seminar, Rosenheim, Germany. 5. Gottlöber, C.; Oertel, M.; Petrak, A.; Wagenführ, A. (2009) Development of milling tools with extreme inclination angle for planing of wood. Proceedings of the 19th International Wood Machining Seminar, Nanjing, China. 6. Gottlöber, C.; Oertel, M.; Wagenführ, A. (2010) Der Einfluss extremer WerkzeugNeigungswinkel beim Umfangsplanfräsen von Holz – Teil 1: Theorie und Erkenntnisstand. holztechnologie, 51 (5): 27-33. 7. Gottlöber, C.; Oertel, M.; Wagenführ, A. (2010) Der Einfluss extremer WerkzeugNeigungswinkel beim Umfangsplanfräsen von Holz – Teil 2: Experimentelle Untersuchungen. holztechnologie, 51 (6): 28-34. 8. Darmawan, W.; Gottlöber, C.; Oertel, M.; Wagenführ, A.; Fischer, R. (2011) Performance of helical edge milling cutters in planing wood. European Journal of Wood and Wood Products, online first (11 January 2011), DOI: 10.1007/s00107-010-0517-8.

29


Investigation of Laser Beam Cutting on Lightweight Sandwich Panels Delenk, Hubertus1; Herold, Jan1; Linde, Hans-Peter2; Gottlöber, Christian1 and Wagenführ, André1 1

2

Technische Universität Dresden, 01062 Dresden, GERMANY Berufsakademie Sachsen, Staatliche Studienakademie Dresden (University of Cooperative Education), Heideparkstraße 8, 01099 Dresden, GERMANY E-mail: [email protected]

ABSTRACT Today lightweight sandwich panels are used for several interior applications. These panels consist of a paper honeycomb core and thin skins made of wood based products (e.g. chipboard or fibreboard). For cutting, milling and drilling the panels are machined by conventional cutting tools. The obtained chips must be disposed with an expensive suction system. Furthermore the cutting process leads to a waste of valuable material. Due to increasing prices for energy and material the investigation of alternative cutting processes may be advantageous. The technology of cutting material with a laser beam is comparatively novel and offers advantages compared to conventional cutting processes. For this reason different experiments have been conducted. For testing the laser beam cutting of lightweight sandwich materials a 25 mm thick panel with a corrugated cardboard honeycomb core has been chosen. The face sheets consisted of high density fibreboard (HDF) with a thickness of 3.2 mm. To find out advantageous process parameters screenings have been executed using a CO2 laser cutting system Eurolaser LCS M-1200 with sealed-off laser technology. Both the cutting of the whole sandwich panel and the cutting of a single skin have been investigated. The maximum feed rate for laser beam cutting of HDF was determined. For the cutting of the whole sandwich panel several processing strategies were developed. On the basis of the results it could be clarified that laser beam cutting in two work steps with changing the machining direction is a favourable solution. Keywords: lightweight sandwich panels; CO2 laser beam cutting; energy reduction; waste reduction

INTRODUCTION Lightweight furniture panels are highly voluminous materials of low density. Today the processing of such is managed by the use of conventional cutting tools. Especially milling tools must have a certain minimum diameter, depending on the material composition and the durability of the tool towards vibrations during the milling process. The reason for research on applying the laser cutting technology on lightweight sandwich panels based on the several advantages which are offered by laser cutting [1]. Compared to traditional cutting processes (for example with a CNC router) laser cutting offers the following advantages: a reduced power consumption of any machine tool [2], narrower kerfs with corresponding material savings, a reduced chip occurrence combined with thusly saved expenses on the suction and reduced maintenance cost due to the lack of wear.

30


STATE OF RESEARCH Laser light is a short-wave monochromatic, electromagnetic, energy-rich and highly ordered radiation. Because of the minimum beam divergence it is possible to focus the laser beam intensely with the aid of optical elements. In this way it is possible to reach highest intensities (intensity = power related to active area) [3]. This physical fact is utilized for material processing with laser beam. The processing result depends on the energy which the work piece absorbs per unit time. Thereby losses occur in form of reflection and transmission. The absorption of the material is not a matter constant – it depends on wave lengths, intensity of the laser radiation, temperature and material thickness. Absorbed laser energy will be converted into heat within the work piece. Heat penetration depth depends on density, thermal conductivity and heat capacity of the material as well as on interaction time with the laser beam [4]. For machining of wood and wood based products a CO2 laser is prior suited. Orech and Kleskeňová [5] have shown light absorption spectra (wave lengths from 0.4 µm to 15 µm) of the wood species mahogany, beech and pine. The investigated materials absorb the light near 10.6 µm (CO2 laser light) of about 90 %. Generally it can be stated that mahogany has the highest and pine the lowest absorption characteristic. This can be explained by the fact that dark wood absorbs light in the investigated range more than light coloured wood [6]. When using a CO2 laser for cutting of non metallic materials the feed rate depends on laser power, material thickness as well as on water and air content of the material. The last two parameters are not constants to describe the process. Water absorbs the radiation of the CO2 laser completely. Therefore high water content leads to a decrease of cutting speed. In contrast air does not absorb the laser radiation. That is why low density wood can be machined easier with CO2 laser radiation [7]. Laser beam machining of organic natural substances causes carbonization of the cutting area on the work piece. With increasing cutting speed and a lower material thickness this effect will be reduced [8]. Regarding the machining of wood with a laser beam it is possible to decrease the usually significant carbonization near the cutting kerf when using an inert gas jet [9].

MATERIALS AND METHODS Test Equipment The tests were performed with the Laser unit Eurolaser LCS M-1200. Because of the design of the machine it is allowed to process materials only with a maximum thickness of 33 mm. Within the laser device a maintenance-free closed CO2 sealed-off laser (Synrad firestar 400) is installed which emits invisible light radiation in the range of about 10.6 µm. By using a pulse rate of 20 kHz the output power reaches a maximum of 400 Watts. Material A Thermopal sandwich panel with a thickness of 25 mm was chosen as sample material. The 3.2 mm thick HDF skins are coated with two melamine resin impregnated papers. The core material, a corrugated cardboard honeycomb, is laminated with a test liner paper. The sandwich panel has a gross density of 340 kg/m³. The storage of the samples and the investigations has been carried out by a temperature of 24 ± 2 °C and a relative humidity of 40 ± 5 %. Specimens with dimensions of 600 mm x 90 mm were used for the tests.

31


Process Parameters The material processing with laser beam is characterized by a large number of influences [6, 10, 11]. These may be initially divided up into parameter groups regarding the laser system, the sandwich panel and the processing environment (see figure 1). Processing environment Medium

Laser system

Climate Machine

Tool

Sandwich panel Work piece

Material

Figure 1: Groups of parameters when cutting sandwich panels with laser beam The different process parameters during the laser beam cutting of sandwich panels can be classified in the corresponding parameter groups cited above. To limit the field of parameters there is no consideration of material properties with respect to the separate components like densities, moistures, structures or used adhesives. Regarding to the emissions no differentiation between gaseous and solid particles was made (compare table 1). Table 1: Process parameters of cutting lightweight sandwich panels using a CO2 laser Parameter Input variable Regulating variable Machine (laser unit) Feed rate vf Focus position zf Tool (laser) Wave lenght λ Focal length of the focusing lens f Mode Pulse rate fp Laser power PL Work piece Thickness s Material Density ρ Moisture ω Temperature TW Medium (processing zone) Emissions Process gas type Processing gas pressure pG Nozzle diameter dN Climate Air temperature TU Rel. Humidity φ

32


Experimental Procedure The investigations were performed with the highest possible feed rate. Therefore the laser power was set to 100 %. Compressed air as process gas and a coaxial nozzle with cylindrical opening duct (diameter 2 mm) as well as a focusing lens with a focal length of 127 mm were used in the experiments. In preliminary tests a contour routing as shown in figure 2 has proven to be suitable. The length of a contour line was 40 mm. Because of starting acceleration of the motor axis drive the laser cut started 5 mm before entering and ends 5 mm after leaving the material (figure 2). Sandwich panel

2

3

Contour routing of the laser tool

1

Figure 2: Contour routing of the tool laser beam for testing a single parameter Initially the cutting of a single skin was investigated. To estimate the maximum feed rate for the laser cutting of 3.2 mm thick HDF with a laser power of 400 Watts results shown in [12] were used. Barcikowski has investigated the laser beam cutting of MDF by using laser power of 500 Watts and 1000 Watts. With this known values power functions were approximated to estimate the feed rate. With the change of coefficients and exponents by bisection the laser power it was possible to find out a corresponding relationship. It was considered that MDF and HDF differ in the density. That is why the determined feed rate was corrected (correction factor 0.83). The result of the estimation is shown in figure 3. 10

Laser power 400 W, HDF

Feed rate [m/min]

8

(3.2,7.8)

6

4

(6.4,3.5)

2 (25.0,0.7) 0

0

5

10

15

20

25

30

Work piece thickness [mm]

Figure 3: Estimation of the maximum feed rate of the cutting on HDF using a laser power of 400 Watts

33


It was tried to use the process parameters for the experiments detailed in [12]. The using of a focusing lens with a focal length of 190.5 mm was not possible. Due to the design of the optical head the use of focusing lenses only with focal lengths up to 127 mm was allowed. Firstly the laser cutting process of the selected sandwich panel was dimensioned without respect to the honeycomb core layer as a 6.4 mm thick HDF (two single layers), respectively as a complete 25 mm thick HDF solid material (compare figure 3). These investigations on both extreme limits (6.4 mm and 25 mm) should supply first knowledge about the range of the useful feed rate on lightweight sandwich panels using a CO2 laser. The cutting depth of the laser tool depends on the feed rate. This is why there is a danger to burn the paper honeycomb core when machining with lower cutting speeds. Figure 4 shows exemplary pictures when starting and finishing the cutting test (compare figure 2). The process was dimensioned for HDF with a thickness of around 25 mm. As shown left in figure 4 the effect of laser beam by entering the sandwich composite is visible up until the middle of the core layer. On leaving the work piece of the laser tool one can see an ignition (figure 4, right). This short phenomenon could be an indication that the burning emissions absorb the laser light. Obviously a corresponding ignition temperature of the medium in the processing zone is predominantly.

Figure 4: Exemplary process pictures of the laser beam cutting on a sandwich panel (tool positions related to the contour routing: left= start, right= finish; variables: vf= 13 mm/s (= 0.78 m/min), zf= -1 mm, fp= 20 kHz, pG= 0.2 MPa) To generate laser beam cutting of lightweight sandwich panels with paper honeycomb core and wood based skins in the following advantageous process variables were determined. In table 2 varied parameters within the possible limits are shown. The variable focus position is synonymously with the distance between laser focus point and work piece surface. Table 2: Variation of process parameters Limits of variaton Step size or intermediate step 1 mm -4 mm < zf ≤ 6 mm 625, 1250, 2500, 5000, 10000 Hz 400 Hz < fp ≤ 20 kHz Pulse rate 100 Hz 400 Hz < fp ≤ 1300 Hz Process gas pressure 0.2 MPa < pG ≤ 0.6 MPa 0.1 MPa Variable Focus position

Using preferred process parameters the maximum feed rate of CO2 laser beam cutting on HDF was determined, followed by the investigation of cutting the whole sandwich panel. Therefore manufacturing strategies were developed. Besides the laser cutting in a single operation, the machining in two processing steps was also investigated. Table 3 shows the field of parameters when cutting sandwich panels in two processing steps. On the basis of the first experiments the corresponding values were determined.

34


Table 3: Field of parameters of cutting sandwich panels in two processing steps during the investigations (MS= manufacturing strategy, PS= processing step; A and B: one processing side, C and D: two directions of machining) Variant 1 2 1 2 fp vf fp vf fp vf fp vf MS PS [kHz] [mm/s] [kHz] [mm/s] [kHz] [mm/s] [kHz] [mm/s] 1.1 13 20.0 19 (1) A 1.1 13 20.0 19 (2) 20.0 109 20.0 19 (1) B 1.1 13 1.1 13 (2) 1.0 13 20.0 19 (1) C 1.0 13 20.0 19 (2) 20.0 109 1.1 13 (1) D 1.1 13 20.0 109 (2)

RESULTS AND DISCUSSION Parameter Transfer It was not possible to transfer the parameter settings of cutting only HDF to the cutting process on sandwich panels. By dimensioning the process in the range between 6.4 mm to 25 mm thickness of HDF there was only a partial cutting of the sandwich panel at the lower face sheet. The feed rate limit was here 12 mm/s (= 0.72 m/min). With further reduction of the feed rate there is a thermal destruction of the paper honeycomb core without a complete cutting of the lower skin. Determination of Preferred Process Parameters By varying the process parameters the preferred settings were determined. It is not advantageously to place the focal point directly into the sandwich material. Choosing a focus position of 4 mm allows a prevention of burning the paper honeycomb core. With an increase of the focus position greater than 4 mm there are poorer process conditions once again. A process gas pressure of 0.6 MPa leads to less burning of the core layer than lower gas pressures. Furthermore there are less burning residues of the laser process on the material surface. Obviously more emissions will be removed from the interaction zone between laser beam and work piece when a higher process gas pressure is chosen. Laser Beam Cutting of HDF The maximum feed rate of cutting HDF (thickness of 3.2 mm) with laser beam was determined to 109 mm/s (= 6.54 m/min). For this experiment the preferred process parameters cited above and a pulse rate of 20 kHz were chosen. Cutting Sandwich Panels with Laser Beam in One Processing Step The cutting of the investigated sandwich panel with CO2 laser beam in a single operation is not possible. As to be seen in figure 5 the partial cutting of the lower face sheet depends on the chosen feed rate. The available energy was too low to cut the double walls of the honeycomb core

35


and the skin zones located below completely. It is assumed that there are energy losses of the laser light into the sandwich core layer. Emissions might to absorb and scatter laser radiation. By choosing a feed rate of 8 mm/s there was a thermal destruction of the paper honeycomb.

Figure 5: Results of cutting the whole sandwich panel with laser beam in a single operation using preferred process parameters (variables: vf= 8, 12 and 16 mm/s; zf= 4 mm, fp= 20 kHz; pG= 0.6 MPa) Cutting Sandwich Panels with Laser Beam in Two Processing Steps The machining strategies A and B in its two variants were investigated (see table 3). The above cited problem with the double cell walls of the honeycomb core is also present here. There is no possibility to cut the whole sandwich panel in two machining steps using the constant orientation of the work piece on the machine table. In contrast to this the variant 1 of the strategy D was proven to be suitable for cutting the chosen sandwich panel in two processing steps. At first the upper skin was machined by laser beam. Following the turn of the specimen the rest of the material was removed using a lower feed rate of the laser tool (see figure 6).

Figure 6: Laser beam cutting of the whole sandwich panel in two processing steps with changing the direction of machining: manufacturing strategy D, variant 1 (variables: PS (1): vf= 109 mm/s (= 6.54 m/min), fp= 20 kHz; PS (2): vf= 13 mm/s (= 0.78 m/min), fp= 1.1 kHz) The left part of the figure 6 shows that the burning residues of the laser process left the cutting kerf opposite to the feed direction of the optical head. Especially at machining points of feed direction changes there is a corresponding deposition of burning residues on the material surface. After the second process step one can see a blackening of the HDF cutting area (compare figure 6, right). Moreover there is a darkening of the paper honeycomb core near the cutting kerf.

36


CONCLUSIONS AND OUTLOOK The investigations have shown that it is possible to cut the chosen sandwich panel made of a corrugated paper honeycomb core and HDF skins with the CO2 laser beam. Therefore two processing steps are necessary. Following the first working step the machining direction was changed (turn of the specimen). Then the second machining operation was carried out. Compressed air as process gas in the tests exclusively was used. Characteristics of the CO2 laser machining process are the carbonization of the HDF cutting areas as well as the darkening of the core layer. There is a need to clarify whether the laser process executed in two machining steps still has advantages related to the power consumption of the machine tool in comparison to a CNC router. When using inert process gas probably the change in colour of the machined surface on HDF and of the honeycomb core can be minimized. Accordingly the danger of burning the core layer and a deposition of burning residues on the material surface would be reduced.

REFERENCES 1.

Delenk, H. (2010) Untersuchung zum Laserstrahltrennen von leichten Sandwichplatten. Diploma thesis, Technische Universität Dresden.

2.

Linde, H.-P.; Siebrecht, D. (2010) Laserstrahl- und Wasserstrahlschneiden – eine Alternative zur spanenden Bearbeitung von Holz- und Holzwerkstoffen. Proceedings of 14. Holztechnologisches Kolloquium, Dresden.

3.

Klocke, F. (2007) Fertigungsverfahren 3: Abtragen, Generieren, Lasermaterialbearbeitung. Springer-Verlag, Berlin Heidelberg.

4.

Poprawe, R. (2005) Lasertechnik für die Fertigung: Grundlagen, Perspektiven und Beispiele für den innovativen Ingenieur. Springer-Verlag, Berlin Heidelberg.

5.

Orech, T.; Kleskeňová, M. (1975) Laser als Bearbeitungswerkzeug in der Holzindustrie. Drevo, 30 (11): 324-326.

6.

Wust, H. (2005) Die Wirkung von Laserstrahlung auf strukturelle, chemische und physikalische Eigenschaften von Holz. Dissertation, Technische Universität Dresden.

7.

Powell, J. et al. (1987) CO2 laser cutting of non-metallic materials. Proceedings of 4th International Conference Lasers in Manufacturing, Birmingham: 69-82.

8.

Treiber, H. et. al. (1990) Der Laser in der industriellen Fertigungstechnik. Hoppenstedt Technik Tabellen Verlag, Darmstadt.

9.

Chryssolouris, G. (1991) Laser Machining: Theory and Practice. Springer-Verlag, New York Inc.

10.

Trasser, F.-J. (1991) Laserstrahlschneiden von Verbundkunststoffen. Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen.

11.

Liebelt, S. (1998) Analyse und Simulation des Laserstrahlschneidens von Faserverbundkunststoffen. Dissertation, Technische Universität Berlin.

12.

Barcikowski, S. (2004) Laserstrahltrennen von Werkstoffen aus Holz. Dissertation, Universität Hannover.

37


38

3. Saw-blade Characteristics Oral Presentations


Nonlinear Vibration of Clamped Saws Ramin M. H. Khorasany and Stanley G. Hutton Department of Mechanical Engineering The University of British Columbia Vancouver, B.C., Canada, V6T 1Z4

ABSTRACT In the present paper, experimental studies were conducted in order to measure the frequency response characteristics of rotating disks both in an idling mode and when subjected to a space fixed lateral force. The applied lateral force (produced by an air jet) was such as to produce displacements large enough that non linear geometric effects were important in determining the disk frequencies. Experiments were conducted on a thin annular disk with the inner radius clamped to the driving arbor and the outer radius free. The results of these experiments are presented with an emphasis on recording the effects of geometric nonlinearities on lateral frequency response. The experimental results show that in the case where significant disk displacements are induced by a lateral force, the frequency characteristics are significantly influenced by the magnitude of forced displacements. In particular, the prediction of linear theory that a zero frequency resonance occurs at so called critical speeds is not valid for the case of disks subjected to large displacements.

EXPERIMENTAL SETUP A schematic of the experimental setup is shown in Figure 1. Four space fixed displacement probes were used to measure the displacement at four different points located around the rim of the spinning disk. Data was collected at a rate of 2 kHz. In order to find the dc level of the disk displacement, the deflection for each probe was averaged over a period of two seconds. Different levels of disk deflection were induced by the application of an air jet whose pressure could be varied to provide different levels of displacement (and nonlinearity). The air jet was located on the opposite side of the blade to probe 3. The magnitude of the air jet excitation was quantified by the steady disk deflection w0 produced at the location of the air jet, when the air jet was applied to the stationary disk. Subsequent disk displacements were measured from this displaced configuration. In order to optimize modal excitation an electromagnet was also used to provide white noise excitation over the frequency range 0-100Hz. Such excitation also had the effect of applying a small dc component of force to the disk. The electromagnet was attached to the frame that supported the driving arbor. Results were obtained by measuring (with reference to the deflected configuration) the displacements of the disk at the location of four probes as the speed was run up from 0 RPM to 4,000 RPM at a constant acceleration rate over a time interval of 500sec and run down from 4,000 RPM to 0 RPM at the same constant rate. Three different levels of air jet excitation were applied to each rotating disk whereby providing different levels of disk deflection and non linear

39


behaviour. The testing sequence was: i) apply air jet excitation to the stationary disk; ii) measure the dc response of the stationary disk at location of the air jet; iii) apply white noise electromagnetic excitation; iv) run disk up to 4,000RPM while continuously measuring the displacements at the four probe locations; v) pause at 4,000RPM and run speed down to 0RPM again measuring the displacements. Using National Instrumentation software “Signal Express” the data was analyzed to determine mean displacement and to produce waterfall and frequency colour maps of the power spectrum that illustrated the variation of disk frequencies with rotational speed. The speed resolution was 25RPM. DC levels of displacements were evaluated with respect to the dc displacement of the stationary disk. The inner diameter of the disc was clamped to the arbor by a 0.1524 m collar and the outer diameter was free. The disk used in this test was typical of those used as wood cutting saw blades. As such they were not perfectly flat. The disk dimensions and flatness indicators are given in Table 1. “Mean dish” refers to average displacement measured at the rim perpendicular to a plane passing through the disk in the clamp, and runout refers to the maximum difference between the mean value and the value at any other point on the disk rim. Table 1. Disk Dimensions and Flatness Indicators Disk #1

Inner Dia.

Outer Dia.

Thickness

Mean Dish

0.1524 m

0.4318 m

1.27 mm

0.15 mm

Figure 1. Experimental Setup

40

Runout 0.13 mm


EXPERIMENTAL RESULTS The intent of the testing conducted in this study was to determine the effect of large disk deformations upon the fundamental behaviour of rotating disks operating in air. Thus the experimental results, conducted in air, will provide the required information. From a fundamental understanding and analysis viewpoint, it is also important to determine if aeroelastic effects play a significant role in the recorded behaviour. The results presented in the Appendix address this issue and conclude that for the blade examined such effects were not significant. Figure 2 presents colour maps of the frequency response of the disk for different levels of applied air jet force for both run up and run down over the speed range dc to 4,000RPM. Mode (n,m) refers to a mode with n nodal circles and m nodal diameters, respectively. w0 is the mean deflection of the stationary disk at the location of the air jet caused by the application of the air jet. The colour bar next to Figure 2a defines the amplitude (dB rms, reference 25.4 mm) which is the same for all of the colour maps shown in this paper. Figure 3 shows the dc amplitude response of the spinning disk versus speed. The amplitude is measured at the location of probe 3. The dc amplitude is calculated using the mean value of two seconds of displacement data. The disk deflection is measured from the position of the stationary disk after the application of air-jet excitation and in these graphs increasing displacement corresponds to the disk moving in the direction of the applied force. From Figure 2a-b it may be noted that when only white noise excitation is used, the frequency response shows behavior which is characteristic of a linear system. The frequencies of backward traveling waves decrease smoothly and reach a speed where the measured frequency is zero or close to zero. This speed is the critical speed of that mode. The critical speeds of the (0,2), (0,3) and (0,4) modes lie between 2,800 and 3,400 RPM. However, the response at these low frequencies is noisy and it is difficult to make definitive statements about the frequency paths in this region. Although air jet excitation is not used for Figure 2a-b there will a small dc component due to the magnetically induced white noise excitation. This is the likely cause for the backward traveling waves of the (0,2) and (0,3) modes that maintain a constant level for the speed range 2800 to 3200 RPM. In Figure 3a, (the run-up case with no applied air jet force) it may be noticed that several stationary waves develop and then collapse in the critical speed region between 2800 and 3200 RPM. Although the disk deflection is small at sub-critical speeds, at supercritical speeds the ratio of dc deflection to thickness approaches 0.37. At such levels, nonlinear effects are important. However, in this case, the level of the applied force is insufficient to sustain the developed stationary waves over a significant speed range. Figure 2c and 2d show the response when w0 h = 0.1 . In these two cases the measured frequencies, at sub-critical speeds, are essentially the same as those measured in the case with only white noise excitation. A difference occurs when the speed approaches the first critical speed around 2,800RPM. At this speed, the measured frequencies of the backward traveling waves of the (0,2) and (0,3) modes level off and have a constant frequency. From Figure 3b it can be seen that in the run-up case, 2,800 RPM also represents the onset of development of a stationary wave. The instability that occurs at this speed causes the amplitude of the disk response to grow. From Figure 3b it may be noted that the amplitude of the dc displacement fluctuates at speeds above 2,800RPM. This is due to the fact that the level of nonlinearity is relatively small and the

41


stationary waves cannot be sustained over a large speed range and collapses shortly after development. When the stationary wave associated with one of these modes collapses the stationary wave for the next mode starts to develop after a short increase in speed. From the dc amplitude response plotted in Figure 3b, it can be seen that in the run up case when the speed approaches 3800 RPM, the last stationary wave becomes unstable and collapses. As shown in Figure 2c , the run up case, in the speed range of 2800 to 3200 RPM a lock-in phenomena in the measured frequency of the (0,0) mode occurs. After this speed range, when the last stationary wave collapses, a jump in the measured frequency of the (0,0) mode is evident. From Figure 3b, it can be seen that in the run down case, the stationary wave starts to develop at a speed close to 3600 RPM. At this same speed a drop in the frequency of the (0,0) mode is also evident in the run down case shown in Figure 2d,. The frequency of this mode then remains almost constant until the speed reaches 2800 RPM. In the present tests, instabilities are observed, at low levels of applied force, coinciding with the initiation of a stationary wave and with the collapse of this wave at higher disk speeds. These instabilities are not aerodynamically induced, but are associated in a nonlinear manner with disk critical speeds. In Figure 2e and 2f the measured frequencies are plotted for the case where w0 h = 0.4 . In this case the measured frequencies of the forward and backward traveling of the (0,2) mode at zero speed are found (by extrapolating the curves to zero speed) to be more separated than the cases with no air jet excitation or when w0 h = 0.1 . The separation of these frequencies for the (0,1) mode is not so significant. Application of the air jet destroys the axial symmetry of the disk resulting in separation of the zero speed frequencies. Hutton et. al. [1] showed that, in the linear case, adding space fixed springs to a rotating disk system, destroys disk symmetry and results in separate backward and forward traveling wave frequencies of the same mode at zero speed. Another source of separation in the measured frequencies is the presence of a non symmetric stress distribution in the disk (due to the applied force). Table 2 presents estimates of the measured frequencies of the (0,2)f and (0,2)b modes of the stationary disk as a function of initial disk displacement produced by the air jet (‘f’ and ‘b’ stand for forward and backward traveling waves, respectively). The frequency values presented in this table correspond to the run-up case. The difference between the measured frequencies of the backward and forward traveling waves of the (0,2) mode increases as the initial displacement due to the air jet excitation increases. As may be noted from Figure 2e and 2f, at sub-critical speeds, the measured frequencies are close to those in Figure 2a which have essentially linear characteristics. The backward traveling waves of the (0,2) and (0,3) modes decreases smoothly when the speed increases until 2,800 RPM. At around 2,800 RPM, a stationary wave starts to develop. After this speed, the measured frequencies of the backward traveling waves of the (0,2) and (0,3) modes maintain a constant level up to 4,000 RPM. The frequencies of the backward traveling waves of the (0,2) and (0,3) modes do not approach zero.

42


Since there is no sign of a sudden drop or change in the frequency and amplitude responses, it can be concluded that in this case, for the range of speed used, there is no secondary instability. Not only do the frequencies of the (0,2) and (0,3) modes maintain a constant level, the frequency curves for higher frequency modes also diverge from their linear characteristics. For example, it may be clearly seen that the frequency response of the (0,6) mode levels off and maintains a constant level. This behaviour occurs concurrently with the development of the stationary wave. The dc configuration of the rotating disk now has characteristics different from those possessed by the disk prior to development of the stationary wave. When the speed decreases from 4,000 RPM in the run down case, the dc amplitude and measured frequencies, follow almost the same path as measured in the run up case. This is because the developed stationary wave is stable at all speeds below 4,000 RPM. Figure 2g and Figure 2h shows the measured frequency response of the disk for the case of w0 h = 0.6 . Once again, the measured frequencies of the (0,2) and (0,3) modes are almost constant in the speed range over which the stationary wave is developed. There is not a significant difference between this case and the previous one with w0 h = 0.4 . Figure 2e-h clearly indicate that no backward travelling waves approach a zero frequency level. Instead, at speeds close to the lowest critical speed they veer off to adopt constant levels independent of speed. Thus, in this sense, no critical speeds exist for these conditions. Table 2- Estimated Frequencies at Zero Speed (Hz); (0,2) Mode (0,2)b

(0,2)f

Difference (Hz)

w0 h = 0.0

65.5

66.2

0.7

w0 h = 0.1

66.4

68.1

1.7

w0 h = 0.4

67.3

70.2

2.9

w0 h = 0.6

67.7

70.9

3.2

43


Figure 2. Frequency Response of Disk #1 for Different Force Levels Run-Up: (a) w0 h = 0.0 , (c) w0 h = 0.1 , (e) w0 h = 0.4 , (g) w0 h = 0.6 Run-Down: (b) w0 h = 0.0 , (d) w0 h = 0.1 , (f) w0 h = 0.4 , (h) w0 h = 0.6

44


Figure 3- Disk 1- DC Displacement versus Speed (probe 3) (a) w0 h = 0.0 , (b) w0 h = 0.1 , (c) w0 h = 0.4 (d) w0 h = 0.6

Figure 4- Disk 1- DC amplitudes at different angular locations for two different configurations of the probes when w0 h = 0.4

45


In order to establish the spatial configuration of the stationary wave, tests were conducted with probes in different locations and with the air jet at 90 degrees (see Figure 4). The angular position of the probe is measured from the x axis (Figure 1) in the counter clockwise direction. It was determined that the stationary wave being excited had (0,3) nodal diameter characteristics. This nodal configuration coincides with that of the mode with the lowest linear critical speed

CONCLUSIONS Experimental results have been presented that illustrate the nonlinear vibration behavior of uniform, thin, spinning disks subjected to a constant spaced fixed lateral force. Different magnitudes of force were applied together with low level magnetic white noise over the frequency range 0-100Hz. Frequency and amplitude characteristics are measured as a function of speed and applied force. Using different levels of air-jet excitation, different levels of nonlinearity are induced and the resulting responses recorded and compared to linear behaviour. DC displacement results obtained for speeds below the lowest critical speed were not sensitive to the level of nonlinearity. In the region of, and above, the critical speed the results obtained illustrate the important effects of nonlinearity. At very low levels of nonlinearity the dc response in the critical speed region was unstable and characterized by the formation and collapse of stationary waves over a short speed interval. Above a certain level of nonlinearity, stationary waves developed and were stable up to the maximum speeds tested. Superimposed on these stationary waves were (smaller) waves corresponding to free vibration modes. The formation of these stationary waves coincided with a change in the path of the frequencies of the backward travelling waves of all the modes recorded. At the speed corresponding to the formation of the stationary wave, existing backward travelling waves veered from their linear path to one where the frequency remained constant with speed. Thus the effect of the nonlinearity was to modify the backward traveling waves such that they do not experience a zero natural frequency at any disk speed. A consequence of the backward travelling waves adopting a constant value is that there are no reflected waves and thus no supercritical speed conditions for flutter instability arise. Although frequency locking has been observed before in rotating disks, it does not appear that this characteristic has been previously related to the presence of a stationary wave.

REFERENCES [1] Hutton, S.G., Chonan, S., and Lehmann, B.F., 1987, “Dynamic response of a guided circular saw,” Journal of Sound and Vibration, 112, pp. 527-539.

46


Practical Measurement of Circular Saw Vibration Mode Shapes Gary S. Schajer 1, Mats Ekevad 2 and Anders Grönlund 2 1

Dept. Mechanical Engineering, University of British Columbia, Vancouver, Canada

2

Division of Wood Technology, Luleå University of Technology, Skellefteå, Sweden

ABSTRACT Natural frequency measurement provides a convenient quantitative method for monitoring the tensioning state of a circular saw. However, it can often be difficult to interpret the measurements because the corresponding vibration mode shapes are not explicitly known, especially when adjacent natural frequencies are close together. A mode shape identification method is presented. It involves using two vibration sensors, one fixed and one orbiting the sawblade circumference. The performance of a simple prototype measurement device using this technique is described. The device could successfully identify the nodal diameter and nodal circle numbers of the major sawblade vibration modes.

INTRODUCTION Dynamic stability is an important factor that influences the cutting accuracy of a circular saw. If a circular saw operates at a rotation speed near to a critical speed, it becomes unstable and cuts inaccurately [1,2]. Circular saws are typically “tensioned” before being put into service to improve their dynamic stiffness and cutting accuracy [3,4]. The tensioning is typically done by rolling [5], heating [6] or hammering [7]. The purpose is to induce residual stresses into the sawblade. The circumferential stresses along the tooth line are tensile, hence the name “tensioning”. The presence of these stresses stiffens the cutting edge and increases the sawblade natural frequencies and critical speed. The amount of the tensioning induced is controlled so that it is large enough to move the sawblade critical speed significantly away from the operating speed, but not so large as to cause sawblade buckling, also called “dishing” [8]. A commonly used industrial method for monitoring the amount of tensioning induced in a sawblade involves inspection of the light-gap between the sawblade surface and the edge of a slightly curved ruler [9]. While this method is convenient and easy to use, it has the drawback that the light-gap observations are subjective and rely on the judgment of the operator. To address this issue, natural frequency measurement has been introduced as a method for saw tensioning evaluation [10,11]. Natural frequencies can readily be measured using a vibration sensor connected to a frequency spectrum analyzer. With modern equipment, the data acquisition and processing system can be as simple as a laptop computer with microphone, sound card and spectrum analysis software. A limitation of the natural frequency measurement method is that is provides numerical values for the frequencies, but it does not identify the corresponding mode shapes. The lack of the mode shape information seriously impedes interpretation of the natural frequency data. The mode shapes can be identified using the Chladni method [12,13], but the measurement procedure is time consuming and unsuited to industrial use. Another possible approach is to

47


compute the expected natural frequencies corresponding to the various mode shapes, and to seek to identify these frequencies from among the measurements. However, it often happens that some natural frequencies are closely spaced, and so it can be difficult to match the measurements to the calculated frequencies and mode shapes. This paper describes an extension to the natural frequency measurement method that identifies both the sawblade natural frequencies and their mode shapes. It involves using two vibration sensors, one of them stationary, and one orbiting around the sawblade circumference while making measurements. A simple laboratory prototype system was designed and built, and its performance for sawblade natural frequency and mode shape identification is described.

THEORETICAL BACKGROUND The vibration modes of a circular saw are defined in terms of their number of nodal diameters and nodal circles. Figure 1 illustrates the mode shapes. In general, the natural frequencies increase with increase number of nodal diameters and circles. The lowest frequency modes (nodal diameter = 0-5 and nodal circle = 0-1) influence cutting behavior [1,2], while the higher frequency modes influence production of aerodynamic noise [14].

Figure 1. Vibration mode shapes of a circular saw, showing 0-3 nodal diameters and 0-1 nodal circles. For standing wave vibration of a symmetric circular saw, without slots or guides, the vibrational displacement is: u (r,θ,t)

=

A cos nθ f(r) cos ωt

(1)

where f(r) is the radial vibration profile, n is the nodal diameter number, θ is the angular coordinate, ω is the angular natural frequency, and A is the complex amplitude (magnitude and phase) of the sawblade vibration. Figure 1 schematically illustrates typical vibration mode shapes. A similar set of vibration modes with sin ωt replacing cos ωt also exists, with nodal lines midway between those shown in Figure 1. For a circularly symmetric structure like a circular sawblade, the two sets of mode shapes differ only in their positions around the sawblade. They have the same radial profiles f(r) and the same natural frequencies.

48


orbiting sensor

θ fixed sensor

sawblade

Figure 2. Schematic diagram of the mode shape identification apparatus. Figure 2 shows the proposed measurement arrangement. It uses two vibration displacement sensors, one at a fixed position, and the second orbiting around the sawblade circumference. The second sensor is attached to an arm pivoted at the center that can be rotated at a constant speed. For a sawblade that is clamped at the centre, the radius of the fixed sensor is approximately midway between the clamp radius and the outer radius. For a free sawblade, the radius of the fixed sensor is close to the inner edge. The orbiting sensor is set so that it measures near the outside edge. These choices of radii are chosen to facilitate distinction between mode shapes with zero and one nodal circles. The two sensor positions will be on opposite sides of the nodal circle line for 1-nodal circle modes, but on the same side for 0nodal circle modes. The relative phase of vibration is opposite for the two cases. During measurements, the orbiting sensor rotates at a constant speed Ω, typically 1 – 5 rad/s. Data are gathered during a complete orbit starting at θ = 0. Under these conditions, θ = Ωt, and Equation (1) becomes: Fixed sensor:

uf (t)

=

Af cos ωt

Orbiting sensor:

uo (t)

= Ao cos nΩt cos ωt

(2) (3)

where Af = A f(rf) and Ao = A f(ro) are the complex vibration amplitudes measured by the two sensors. The natural frequency ω can be determined by Fourier analysis of the signal from the fixed sensor, uf (t) [15]. The nodal diameter number n can similarly be determined the signal from the orbiting sensor, uo (t). Multiplying Equation (3) by cos mθ = cos mΩt gives: uo (t) cos mΩt = Ao cos nΩt cos mΩt cos ωt = ¼ [ Ao cos ( (ω + (n+m)Ω)t ) + Ao cos ( (ω – (n+m)Ω)t ) + Ao cos ( (ω + (n–m)Ω)t ) + Ao cos ( (ω – (n–m)Ω)t ) ]

49

(4)

(5)


Equation (5) shows that uo (t) cos mΩt has Fourier components at frequencies ω ± (n±m)Ω. A Fourier component at frequency ω exists only when m = n. Thus, the nodal diameter n can be found by finding the Fourier component of the signal uo (t) cos mΩt for series of values of m = 0,1,2,... In theory, a non-zero Fourier component exists only for the case where m = n. In practice, non-constant signals and the presence of noise cause small numerical values to be evaluated for all m. However, the numerical value expected with m = n should be much larger than with m ≠ n. Thus, the desired m = n value is identified as the one that maximizes the absolute value of the Fourier component, ½ Ao, at frequency ω. The relative phase of the vibrations measured at the two sensors can be determined from the phases of the corresponding Fourier components. Taking the ratio gives: R = Ao / 2Af

(6)

All quantities is Equation (6) are complex quantities. The angle of the ratio R corresponds to the phase difference between the two sensor signals at frequency ω. When the angle is close to 0, the signals are in-phase, indicating a vibration mode with 0 nodal circles. When the angle is close to π, the signals are out-of-phase, indicating a vibration mode with 1 nodal circle. If desired, the higher nodal circle modes can be identified by specific radial placement of the orbiting sensor. These vibration modes can also be identified by their much higher frequencies.

METHOD The proposed natural frequency and mode shape identification method comprises the following steps: 1.

Mount the sawblade in the apparatus schematically shown in Figure 2 and adjust the radius of the orbiting sensor to be near to the outside edge, and the radius of the fixed sensor to be midway between the inner and outer edges for a clamped saw, and near the inner edge for a free saw.

2.

Strike the sawblade with a soft-faced hammer at a point along the same diameter as the stationary sensor. Such an impact tends to excite symmetrical vibration modes of the type shown in Equation (1). The use of a soft-faced hammer tends to excite the lower frequency vibration modes that are of interest here.

3.

Allow a few seconds for the initial transient vibration to decay and then record measurements from both sensors while the orbiting sensor completes one revolution at constant speed. If is convenient for the subsequent data analysis to adjust the sampling rate and rotation speed such that the number of measured data is a power of 2, for example 1024, 2048 or 4096.

4.

Determine the sawblade natural frequencies ω by fast Fourier analysis of the data from the stationary sensor.

5.

Evaluate Equation (4) for values of m = 0, 1, 2, ... For each of the natural frequencies identified in Step 4, identify the value of m that maximizes the Fourier component at frequency ω. This value of m equals the nodal diameter number n for that natural frequency. If the angle of the complex ratio R in Equation (6) is close to 0, the nodal circle mode = 0. If close to π, the nodal circle mode = 1.

50


EXPERIMENTS A simple laboratory prototype sawblade test stand was constructed according to the arrangement shown in Figure 2. Since most vibration frequencies of interest are in the audio range, audio microphones were chosen as convenient low-cost vibration sensors. These were mounted near to the surface of the sawblade, one in a fixed position below the sawblade, and one attached to a rotating arm above so that it could orbit around the sawblade circumference. The phase of the signal of the orbiting microphone was inverted so that the signals from the two microphones correspond to the same side of the sawblade. In the simple prototype system, the arm was rotated manually, with one revolution taking about 5s. The microphones were connected to a computer sound card, from which their signals could be recorded using standard sound recording software. The resulting data files in .wav format were read and analyzed by custom software that computes Equation (4) and the associated Fourier analysis, and displays the results. The test saw was 700mm OD, 120mm ID and 2.8mm thick. It was taken from sawmill stock, and was lightly tensioned. Table 1 lists the theoretical natural frequencies and mode shapes of the sawblade, calculated using CSAW software [16]. Figure 3 shows the vibration power spectrum measured using the proposed method. The number above each peak indicates the vibration frequency, the digit below and to right is the identifed nodal diameter number, and the digit to the left is the identified nodal circle number. These results are also listed in Table 1. It can be seen that the measurement method successfully identifies the various mode shapes.

Figure 3. Power spectrum of the test sawblade, showing the identified natural frequencies and mode shapes.

51


Theoretical Results

Measured Results

Frequency Hz

Nodal Circles

Nodal Frequency Diameters Hz

Nodal Circles

Nodal Diameters

20

0

0

-

-

-

24

0

1

-

-

-

44

0

2

42

0

2

85

0

3

88

0

3

142

0

4

144

0

4

165

1

0

-

-

-

176

1

1

-

-

-

210

0

5

211

0

5

218

1

2

-

-

-

291

1

3

297

1

3

298

0

6

288

0

6

384

0

7

375

0

7

410

1

4

417

1

4

489

0

8

471

0

8

Table 1. Theoretical and measured sawblade natural frequencies and mode shapes. In Figure 3, an additional frequency peak appears at 200 Hz. This is a multiple of the 50Hz line frequency that was present as a background noise during the measurements. Although not appearing in Figure 3, frequency peaks at 50, 100 and 150 Hz occasionally appeared in other test measurements. The associated “mode shapes” of all such background noise sources are identified as having 0 nodal diameters because the associated signal is constant. The corresponding nodal circle number is identified as 1 because the phase inversion applied to the rotating microphone makes an in-phase signal from a noise source appear out-of-phase. The various vibration peaks in Figure 3 have greatly differing sizes, some very large and some very small or even absent. The relative sizes of the peaks depend on the radial position of the impact used to excite the sawblade vibration, the hardness of the hammer used, and the time interval between the impact and the measurements. Some vibration modes, for example the 0-3 mode, were always large and of long duration, while others such as the 0-4 mode were smaller and decayed rapidly. Several vibration modes with one nodal circle decayed very rapidly and could not be observed easily. In Figure 3, the 8 nodal diameter vibration mode does not appear, although it was occasionally large enough to be seen in other measurements. The manual rotation of the orbiting microphone was a significant source of variability. Irregular rotation speed could cause some of the modes shapes to be incorrectly identified. With practice, it was possible to rotate the microphone fairly smoothly and get consistent results. However, a mechanized rotation system would certainly be preferable. Given the promising results obtained with the simple prototype, plans are in hand to design and build a fully mechanized measurement system rugged enough for industrial use. To be effective in a 52


noisy environment, it is planned to use non-contact displacement sensors instead of microphones to measure the sawblade vibration. They will also allow measurement of very low vibration frequencies.

CONCLUSIONS Experimental measurements confirm that the proposed circular saw vibration mode shape identification method is effective. These mode shape identifications are useful when seeking to control sawblade tensioning by monitoring the natural frequencies of specific vibration modes, particularly when adjacent vibration modes have similar natural frequencies. The laboratory prototype measurement device was simply constructed using readily available components, with computer microphones used as the vibration sensors, and a computer with sound card to acquire the data collection and analyze the results. There was some variability in the performance of the prototype system due to inconsistencies in the manual method for impacting of the sawblade to excite vibration, and for rotating the orbiting microphone. A mechanized system is planned to provide more accurate and consistent functionality, and to be suitable for industrial use.

ACKNOWLEDGMENTS Author GS thanks the Luleå University of Technology, Sweden, and the Natural Science and Engineering Research Council, Canada, for their support of the research reported here.

REFERENCES 1.

Dugdale, D.S. (1966) Stiffness of a Spinning Disc Clamped at its Centre. Journal of the Mechanics and Physics of Solids, 14:349-356.

2.

Mote, C.D. and Nieh, L.T. (1973) On the Foundation of Circular-Saw Stability Theory. Wood and Fiber, 5:160-169.

3.

Dugdale, D.S. (1966) Theory of Circular Saw Tensioning. International Journal of Production Research, 4:237-248.

4.

Schajer, G.S. (1984) Understanding Saw Tensioning. Holz als Roh- und Werkstoff, 42:425-430.

5.

Pahlitzsch, G. and Friebe, E. (1973) Über das Vorspannen von Kreissägeblättern. I. (Tensioning of circular sawblades. Part I.), Holz als Roh- und Werkstoff, 31:429-436

6.

McKenzie, W.M. (1969) How does Heat Tensioning of Saw Blades Work? CSIRO Forest Products Newsletter, 363:2-3

7.

Barz, E. (1963) Vergleichende Untersuchungen über das Spannen von Kreissägeblattern mit Maschinen und mit Richthämmern. (Comparative studies of tensioning of circular sawblades with machines and by hammering). Holz als Rohund Werkstoff, 21:135-144. CSIRO Australia Translation No. 6583.

8.

Williston, E.M. (1989) Saws: Design, Selection, Operation, Maintenance. Miller Freeman Publications. San Francisco, CA.

9.

Schajer, G.S. (1992) North American Techniques for Circular Saw Tensioning and Leveling: Practical Measurement Methods. Holz als Roh- und Werkstoff, 50:111-116.

53


10. Szymani, R. (1984) Electronic Evaluation of Circular Saw Tensioning. Processing, 9:20-21.

Timber

11. Szymani, R. (1987) Dynamic Design of Saws: from Theory to Practice. Forest Industries 114(3):24-26; World Wood 28(2):40-42. 12. Chladni, E.F. (1787) Entdeckungen über die Theorie des Klanges, Leipzig. 13. Schajer, G.S. and Steinzig, M. (2008) Sawblade Vibration Mode Shape Measurement Using ESPI. Journal of Testing and Evaluation, 36(3):259-263. 14. Dugdale, D.S. (1969) Discrete Frequency Noise from Free Running Circular Saws, Journal of Sound and Vibration, 10(2):296-304. 15. Dalhquist, G., Björk, Å. and Anderson, N. (1974) Numerical Methods, Prentice-Hall, Englewood Cliffs, NJ. 16. Schajer, G. S. (2007) CSAW 4.0 Computer Software for Optimizing Circular Saw Design, Wood Machining Institute, Berkeley, CA.

54


Wood Chip Formation in Circular Saw Blades studied by High Speed Photography Mats Ekevad1, Birger Marklund1, Per Gren2 1

Luleå University of Technology, Division of Wood Science and Technology, Skeria 3, SE931 87 Skellefteå, Sweden 2 Luleå University of Technology, Division of Experimental Mechanics, SE-971 87 Luleå, Sweden [email protected]

ABSTRACT Films of wood chip formation have been captured. The films have been recorded with a high speed camera during rip sawing of wood with a circular saw blade at full speed. The test apparatus consists of a circular saw intended for rip sawing with a blade diameter of 400 mm and a rotational speed of 3250 rpm. A specially designed saw blade with only 4 teeth was manufactured in order to get a relatively large chip thickness with a low feeding speed. The 4 teeth were made with rake angles of 0°, 10°, 20° and 30° in order to ascertain the influence of different rake angles. A constant feeding speed system was used during the tests. Wooden boards were cut along the side so that the camera could record the cutting sequence without any interference of material between the cutting teeth and the camera. A transparent plastic sheet was placed as close as possible to the cutting teeth, in-between the cutting teeth and the camera. The cuts were made so as to not go through all of the height of the samples. Tests were made for green, dry and frozen green pine boards, for both counter cutting and climb cutting cases. In addition, some Mozambican wood species were cut. The films, recorded at 40000 frames/s, show the cutting sequence along the trajectory of the tooth in question and the creation of the wood chip. Details such as the compression of the wood chip in the gullet, the movement of the wood chip inwards and outwards in the gullet and finally the exit out from the gullet are visible. The chip size and chip movement depend strongly on the rake angle and on whether the wood is green, dry, frozen or unfrozen.

INTRODUCTION Digital high speed cameras are increasingly being used to capture high speed events. Today there is no problem filming the wood cutting process at the tip of a saw tooth in a circular saw blade at full speed. This was also possible with the older analogue film techniques, but at a greater effort and cost. Indeed, Grönlund [1] shows a sequence of pictures of saw dust movement in circular saw gullets recorded with a high speed camera in 1954. The pictures are taken from Englesson et al. [2]. The conclusions from these pictures were that counter cutting with a rake angle of 27° gave a favourable saw dust transport in the gullet. However, climb cutting with the same rake angle resulted in saw chip movement in-between the gap between the preceding saw tooth and the work piece. This produced unfavourable heat generation. The solution to the problem was to change to a rake angle of -7° where climb cutting gave a favourable saw chip transport down into the gullet. Franz, in 1958 [3], describes a flash photography technique to take single pictures during a high-speed cutting process, but writes that “it was not possible to photograph a sequence on the removal of a single chip”. He shows

55


and defines chip formation types from the pictures. References [4, 5, 6] discuss chip formation processes and use photography techniques to study chip formation. This paper shows a method to make films of the chip formation process in a circular saw blade at high speed. Films were recorded when rip sawing 3 species of wood under different conditions. The purpose of the filming was to test the filming method itself, but also to use the films to show the chip formation process and the chip transport in the gullet for counter cutting and climb cutting and for various rake angles, wood species and wood conditions. The filming project was a part of a research project with the goal to increase knowledge on how to increase saw mill yield.

MATERIAL AND METHODS Fig. 1 shows the principles of cutting in the experiments. The saw blade cuts along the side of the workpiece (a board) in order to view the cutting from the side through a glass sheet.

Fig 1. Principle of cutting along the side of a board. Fig. 2 shows the experimental setup with the camera and lighting devices. The camera was a Photron Fastcam SA1.1 from Photron USA Inc., San Diego, California, USA. The settings were 40000 frames/s, exposure time 1/184000 s, picture size 512x256 pixels with 12 bit grayscale resolution and a field of view of 142x71 mm. Each film was about 0.1 s long (real-time). The saw blade was specially designed with 4 cutting teeth with rake angles of 0°, 10°, 20° and 30°, see Fig. 3. Dummy teeth which did not cut were placed before and after the cutting teeth in order to obtain the right shape of the tips of the cutting teeth and the gullets. Counter cutting and climb cutting at two different feed speeds were tested. The gullet feed index (GFI=chip volume/gullet volume) varied between 25% and 33% for the different teeth at low feed speed, and between 42% and 50% at high feed speed. Five different materials were tested, dry pine, green pine, frozen green pine, dry Namuno and dry Ironwood. Table 1 shows the tested combinations and Tables 2 and 3 show the cutting data that were used.

56


Table 1. Tested combinations. Each row is one test. GFI is gullet feed index. Dry pine

Green Pine

Frozen green pine

Namuno

Ironwood

Climb cutting

x

Counter cutting

x x x x x x x x

x x x x x x x x x x x x x x

high GFI

x x x x x x x x x x x x x x x x

x

low GFI

x x x x x x x x

Table 2. Cutting data for low gullet feed index radius (mm) cutting height (mm) gullet area (mm2) cutting speed (m/s) feed per tooth (mm) inlet/outlet angle (rad) outlet/inlet angle (rad) cutting length (mm) mean chip thickness (mm) GFI (%)

Tooth 0° 200 60 195 68.1 0.819 0.795 0 159 0.309 25.2

Tooth 10° 200 60 195 68.1 1.07 0.795 0 159 0.403 32.9

Tooth 20° 200 60 195 68.1 1.05 0.795 0 159 0.396 32.2

Tooth 30° 200 60 195 68.1 0.939 0.795 0 159 0.354 28.9

Table 3. Cutting data for high gullet feed index feed per tooth (mm) mean chip thickness (mm) GFI (%)

Tooth 0° 1.37 0.518 42.2

Tooth 10° 1.62 0.612 49.9

57

Tooth 20° 1.60 0.605 49.3

Tooth 30° 1.49 0.563 45.9


saw blade camera

glass window

lights

Fig.2. Experimental setup. Photography taken through glass window. cutting tooth 0° 30° 10° 20° dummy teeth

Fig.3. Circular saw blade with 4 cutting teeth with rake angles of 0°, 10°, 20° and 30°. Also dummy teeth (not cutting) before and after each cutting tooth.

58


RESULTS The films show slow motion movies of the moving cutting teeth along the cutting length. Chip formation and chip movement in the gullets are visible. Figs. 4, 5, 6 and 7 show photos at one instant in time of chips in the gullets for some of the combinations tested. The tooth that is shown in each figure moves from left to right and slightly downwards for the counter cutting cases (Figs. 4 and 7) and slightly upwards for the climb cutting cases (Figs. 5 and 6).

a) rake angle 10° b) rake angle 30° Fig.4. Chip formation and movement in the gullet. Counter cutting, green pine wood, high GFI

a) rake angle 0° b) rake angle 30° Fig.5. Chip formation and movement in the gullet. Climb cutting, frozen green pine wood, high GFI

DISCUSSION AND CONCLUSIONS Chip formation in general for all cases seems to be of type 2, where the chip is formed close to the tip of the tool [1, 3, 6]. The field of view was not small enough to reveal the exact chip formation at the cutting edge.

59


a) rake angle 0° b) rake angle 30° Fig.6. Chip formation and movement in the gullet. Climb cutting, dry pine wood, high GFI

a) rake angle 0° b) rake angle 30° Fig.7. Chip formation and movement in the gullet. Counter cutting, dry Ironwood, low GFI The chips were broken more or less into small pieces directly after the separation from the wood sample (Figs. 6 and 7) but also long chips were formed (Figs. 4 and 5). For dry wood there were a higher proportion of small pieces than for green wood. Namuno and Ironwood gave especially small chip pieces (Fig. 7). The proportion of chips that were long was high for green wood (Fig. 4), especially for frozen green wood (Fig. 5). Long chips were compressed into a zig-zag form or a curled form in the gullet (especially in Fig. 4a). In one case the long chip twisted and moved in-between the saw blade body below the gullet and the glass window. The chips that were generated moved in a straight line inwards along the rake face or in a straight line with a slight angle in front of the rake face (Fig. 5). When this flow of chips hit the bottom of the gullet, the chips stopped and the chips gathered in the gullet, or the flow bounced outwards and forwards along the bottom of the gullet. Small rake angles gave more of the first type, namely bouncing and forward flow of chips in the gullets (Figs. 5a and 6a). Large rake angles gave more of the second type, namely gathering of chips in the gullets (Figs. 4b and 6b).

60


Counter cutting (Figs. 4 and 7) created more long milling types of chip in the beginning of the cutting length and more small chips at the end of the cutting length. For climb cutting (Figs. 5 and 6) this was reversed. Due to this sequence it seemed that the occurrence of small chips that went into the gap between the blade and the glass wall was more frequent for climb cutting cases than for counter cutting cases. Climb cutting creates a greater proportion of chips in the beginning along the cutting length than at the end. This means that a relatively large amount of chips have to be transported in the gullet a long distance along the cutting length. In the alternative case of counter cutting, a greater proportion of the chips are created at the end along the cutting length than at the beginning. This means that a smaller proportion of chips have to be transported a long way in the gullet compared to the climb cutting case. The result of this reasoning is that the probability of chips getting in-between the saw blade and the glass window may be higher in the climb cutting cases than in the counter cutting cases. Chips going in-between the saw blade and the kerf wall results in heating of the saw blade and can thus destabilise the blade.

ACKNOWLEDGEMENT The authors express their gratitude to the European Regional Development Fund, Objective 2, Northern Sweden, via Tillväxtverket (the Swedish Agency for Economic and Regional Growth) and Vinnova (the Swedish Agency for Innovation Systems) for financial support.

REFERENCES 1.

Grönlund A. (2004) Träbearbetning (in Swedish). Publ nr 0405011. Trätek-Institutet för träteknisk forskning, Stockholm, ISBN 91-88170-32-2.

2.

Englesson T., Hvamb G., Thunell B. (1954) Noen resultater fra undersökelser over sagning med og mot fibrerne (counter- and climb-ripsawing). NTI meddelande nr 7. Oslo, Norway.

3.

Franz N. C. (1958) An Analysis of the Wood-Cutting process. PhD Thesis, University of Michigan Press, Ann Arbour.

4.

Kivimaa E. (1950) Cutting Force in Woodworking. Publ. No 18. The State Institute for Technical Research, Helsinki.

5.

McKenzie W.M. (1961) Fundamental Analysis of the Wood-Cutting Process. Department of Wood Technology. School of Natural Resources, University of Michigan, Ann Arbour.

6.

Koch P. (1964) Wood Machining Processes. Ronald Press, N.Y.

61


62

4. Tool Material and Tool Wear Oral Presentations


Effect of Cutting Tools with Advanced PVD Coating on Reduction of Power Consumption Minami, Toru and Nishio, Satoru KANEFUSA CORPORATION, Ohguchi-cho, Niwa-Gun, Aichi-Ken, JAPAN

ABSTRACT Today, technology has developed to reduce the environmental impact in the wood machining field too. We have commercialized a wide variety of products which use our original PVD coating technology, and these products can achieve a longer lifetime and better cutting quality compared with uncoated tools. In this research, we measured not only edge wear and lifetime of tools but also power consumption to demonstrate the effects of our coating technology in planing on a moulder machine and a CNC machine for timber construction. We have found that the coated HSS planer knives can make the tool life 4 times longer than that of uncoated knives, and reduce power consumption of the spindle by 20%. The cutting quality of the coated knives was much better than that of uncoated knives throughout the lifetime. It was also found that the coated carbide inserts showed more than twice the lifetime than that of TiN coated inserts from a competitor in joints machining for timber construction. A big advantage was that the feed speed of the tool controlled by a cutting load did not slow down because the cutting power required by the coated tool was 40% lower than that of the TiN coated tool. This longer lifetime and lower cutting power results from an ideal cutting edge shape due to the optimized coating properties. It was concluded that reducing environmental impact can be possible as well as a long lifetime and a good cutting quality by using tools with our advanced PVD coating.

INTRODUCTION In the field of wood machining tools, high speed tool steel (HSS) and tungsten carbide are commonly used. Those materials are also coated with hard films such as transition metal nitrides. Cutting tools coated with Chromium nitride (CrN) coatings prepared by the PVD method have been commercialized [1, 2], because CrN coatings have excellent wear and corrosion resistance for cutting wood. Due to increased public interest in environmental problems in recent years, a reduction of power consumption in machining processes by better durability and sharpness of tools is desired. Therefore, the performance of CrN coated cutting tools should be improved, too. Some research on the performance of CrN coated cutting tools for wood and wood based materials has been

63


reported [3, 4, 5, 6, 7]. However, there are not many studies on the performance of CrN coated cutting tools in actual operations, and it is rare that the relationship between cutting performance and cutting power consumption is investigated. In this study, we prepared HSS and tungsten carbide cutting tools coated with CrN based coatings, and they were used in actual wood machining operations at our customers. The aim of this study is to examine the effects of our advanced coatings on wear resistance, cutting quality and power consumption. We have also developed the new coating for cutting wood, and demonstrated its performance.

EXPERIMENTAL DETAILS Sample preparations CrN based coatings were deposited by using a PVD ion plating system. Planer knives made of high speed tool steel (HSS, JIS SKH-9) and tungsten carbide inserts (K grade) were used as a substrate. The size of the planer knives was 150mm long, 30mm wide and 3mm thick. The inserts were 12-30mm long, and the width and the thickness were 12mm and 1.5mm, respectively. The substrates were ultrasonically cleaned in an alkaline solvent, water and an organic solvent. Prior to deposition, substrates were heated to 673K, and sputter cleaned by Ar ions. The coated tools were ground to create sharp cutting edges after the coating process, and the coatings remained only on the rake face of the tools. Cutting tests The uncoated and coated HSS planer knives were tested on a moulder machine (TOPSPEC) at a glued lumber factory. The coating was a CrN based multilayer film. The knives were mounted on a cutter head that was 203mm in diameter with 10 teeth. The jointing procedure was done on the moulder to produce the same cutting circle for all 10 knives in the cutter head. Work piece materials were Spruce, European red pine and Japanese cedar. Revolution of the spindle, feed rate and the depth of cut were 6000RPM, 85m/min and 0.5mm, respectively. We used a clamp-on power meter to measure power consumption of the spindle from the beginning to when the cutter head was taken off for resharpening. Power consumption of the coated knives was compared with that of the uncoated knives. The coated tungsten carbide inserts were tested on a CNC machine (HEIAN) for timber construction at a customer. We used a router bit which was applied for a dove tail joint of wood. The inserts were coated with a CrN film and a CrN based multilayer film. TiN coated inserts which were delivered from a competitor were also used as a comparison. Work piece materials were various solid woods and glued lumber. Revolution of the spindle, feed rate and the depth of cut were 6000RPM, 1-3m/min and 20mm, respectively. The cutting load automatically controlled the feed rate. An operator of the machine estimated the lifetime of the coated inserts by checking the cutting quality. We got the used inserts from the customer to perform an in-house cutting test, and measured cutting power and cut surface quality. Spruce glued lumber was used as a work

64


piece, and cutting conditions were the same as those at the customer. In both cases of HSS planer knives and tungsten carbide inserts, scanning electron microscopy (SEM) was used to observe wear characteristics of the cutting edges, and a surface roughness meter was used to measure the amount of wear and the cross-section shapes of the cutting edges.

RESULTS and DISCUSSIONS Coated HSS planer knives Fig.1 (a) and (b) show SEM micrographs of the cutting edge both of the used uncoated and the coated HSS planer knives. In the case of the uncoated HSS knife, the cutting edge becomes round due to heavy wear. An obvious crater wear on the rake face is also observed. On the other hand, the coated HSS knife still shows a very sharp cutting edge due to the coating on the rake face. Wear on the clearance face can be seen. (a)

(b)

Rake face

Clearance face

Fig. 1 SEM micrographs of the cutting edge, (a) uncoated and (b) coated HSS planer knives after use.

(a)

(b)

Rake face

Clearance face

Fig. 2 Cross-section shapes of the cutting edge, (a) uncoated and (b) coated HSS planer knives after use. The cross-section shapes of the cutting edges both of the used uncoated and the coated HSS planer knives are shown in Fig. 2 (a) and (b). The uncoated HSS knife shows a round cutting

65


edge, and a large crater wear is visible. The amount of edge recession is about 15 microns, and the wear width on the rake face is about 90 microns, while the wear width on the clearance face is small. On the other hand, it is clear that the coated knife still has a very sharp cutting edge, and the amount of edge recession is only 4 microns. The wear width on the rake face is almost nothing, but the wear width on the clearance face is about 120 microns. The appearance of wear on the uncoated and coated knives was very different due to the effect of the coating. The results of power consumption of the spindle versus working time when the uncoated and the coated HSS knives were used are shown in Fig. 3. It is noticed that the power consumption for the coated HSS knives is about 20% lower than that of the uncoated knives. The lifetime of the coated knives was about 96 hours, which was 4 times longer than that of the uncoated knives. The reason for the low power consumption and the long lifetime must be the shape of the sharp cutting edge, as shown in Fig. 3 (b). 600 Power consumption （kWh）

Regrinding

500 Regrinding

400

Regrinding

300 Regrinding

Regrinding

200

Uncoated

100

Coated 0 0

50

100

150

Time (hour)

Fig. 3 Power consumption of the spindle versus working time. Coated tungsten carbide inserts Fig. 4 (a), (b) and (c) show SEM micrographs of the cutting edges of the coated tungsten carbide inserts used at a customer. As clearly shown in Fig. 4 (a), the TiN coating on the rake face was worn out, and the edge became round. It was found that TiN coating had no effect on cutting wood. On the other hand, the CrN coating and CrN based multilayer coating show relatively sharp cutting edges due to the coatings on the rake face. The CrN based multilayer coating had the least wear.

66


(a)

Rake face

(b)

(c)

Clearance face

Fig. 4 SEM micrographs of the cutting edge, (a) TiN coating, (b) CrN coating and (c) CrN based multilayer coating. The cross-section shapes of the cutting edges of TiN, CrN based and CrN based multilayer coatings after use are shown in Fig. 5 (a), (b) and (c). The amount of edge recession of TiN, CrN and CrN based multilayer coatings was 27, 22 and 17 microns, respectively. The wear width on the rake face of TiN, CrN and CrN based multilayer coatings was 18, 11 and 10 microns, respectively. The results clearly showed the effect of keeping the cutting edges sharp for a long period, and the CrN based multilayer coating had the best result. The lifetime of the TiN, CrN and CrN based multilayer coatings was 5, 7 and 10 days, respectively. It is obvious that small edge recession and a sharp cutting edge affected the lifetimes. (a)

(b)

(c)

Rake face

Clearance face

Fig. 5 Cross-section shapes of the used cutting edge, (a) TiN coating, (b) CrN coating and (c) CrN based multilayer coating. Then we performed an in-house cutting test with the used inserts to measure cutting power and cut surface quality. Fig. 6 shows the net cutting power when machining Spruce glued lumber. Cutting power of the TiN coating was the highest, and the lowest was for the CrN based multilayer coating, which was about 40% less than that of the TiN coating. A big advantage was that the feed speed of the tool controlled by a cutting load did not slow down because the cutting power required by the coated tool was 40% lower than that of the TiN coated tool. Fig. 7 (a), (b) and (c) show cut surface quality. The used TiN coated inserts caused fuzzy grains and big chippings of wood in the exit side as indicated by arrows in the pictures, while the inserts coated with the CrN coating and the CrN based multilayer coating showed a much better quality. More specially, the insert with the new coating caused less chipping of the wood. We found that there was a good correlation between the shape of the cutting edge, cutting power and cut surface

67


quality. It was also revealed that the CrN based multilayer coating could keep the cutting edge sharp effectively for a long period of time. 2.5 Competitor TiN coating KANEFUSA CrN based coating KANEFUSA New coating

Cutting power (kW)

2.0 1.5 1.0 0.5 0.0 0

5

10

15

Time (sec)

Fig. 6 Cutting power comparison when machining Spruce glued lumber. (a)

(b)

68


(c)

Fig. 7 Cut surface quality comparison when machining Spruce glued lumber, (a) TiN coating, (b) CrN coating and (c) CrN based multilayer coating.

CONCLUSIONS In this study, we have investigated not only the lifetime and wear resistance but also power consumption of the coated HSS planer knives and tungsten carbide inserts at our customers. The results are summarized as follows. 1. HSS planer knives coated with a CrN based coating showed a 4 times longer lifetime and 20% lower power consumption than that of the uncoated knives. Due to the coating on the rake face, the cutting edge can stay sharp for a long time, and the wear behavior changes drastically. 2. Tungsten carbide inserts coated with a CrN coating and a CrN based multilayer coating showed 1.4 times and twice the lifetime than that of TiN coated inserts. Cutting power of the insert coated with a CrN coating and the a CrN based multilayer coating was 20% and 40% lower than that of TiN coated insert. Compared with TiN coating, our two coatings showed a much better cut surface quality, because the shape of the cutting edge was sharp. 3. It is obvious that appropriate coatings have a great effect on not only increasing lifetime and good cut surface quality but also lowing power consumption. Reducing the environmental impact can be possible by using tools with the advanced coatings.

REFERENCES 1. Soga, K., Kawai, T., Kozawa, Y. (1995) JP Pat. 1941098 2. Tsuchiya, A., Nishio, S., Soga, K. (1997) JP Pat. 2673655 3. Darmawan, W., Tanaka, C., Ohtani, T., Usuki, H. (2000) Wear characteristics of some coated carbide tools when machining hardboard and wood-chip cement board. Mokuzai Kougyou, 55: 456-460. 4. Djouadi, M. A., Nouveau, C., Beer, P., Lambertin, M. (2000) CrxNy hard coatings deposited with PVD method on tools for wood machining. Surf. Coat. Technol., 133-134: 478-483.

69


5. Nouveau, C., Djouadi, M.A., Dece`s-Petit, C. (2003) The influence of deposition parameters on the wear resistance of CrxNy magnetron sputtering coatings in routing of oriented strand board. Surf. Coat. Technol., 174-175: 455-460. 6. Faga, M. G., Settineri, L. (2006) Innovative anti-wear coatings on cutting tools for wood machining. Surf. Coat. Technol., 201: 3002-3007. 7. Nouveau, C., Jorand, E., Dece`s-Petit, C., Labidi, C., Djouadi, M.A. (2005) Influence of carbide substrates on tribological properties of chromium nitride coatings: application to wood machining. Wear, 258: 157-165.

70


Carbide Tipping and Fatigue Strength of Band Saw Blades By F. Scholz, U. Heisel, J. Troeger, M. Großmann, et al. Scholz F., University of Rosenheim, Faculty of Wood Technology and Construction Heisel U., University of Stuttgart, IfW Institut für Werkzeugmaschinen Tröger J., University of Stuttgart, IfW Institut für Werkzeugmaschinen, Großmann M., University of Stuttgart, IfW Institut für Werkzeugmaschinen, Hemer A., Beier R., Sidharta A., Cheng K., University of Rosenheim, Dep. Research and Development

ABSTRACT Band sawing technologies have been improved in two different aspects during the last years:  Development of welding and soldering techniques enabling the production of carbide tipped bandsaws,  Increasing the precision by high pretensioning. Both effects require enhanced fatigue strength of sawblades. The maintenance intervals are assumed to increase 5 times. This yields fatigue cycle numbers which require sawblade materials with fatigue capacities beyond the endurance limit. The situation is getting still more severe when these blades are used on machines with enhanced pretension level. A research program was carried out to assess the different possibilities to improve the fatigue strength. This paper presents a survey of the achieved results. The project was financed by AIF and carried out by the IfW Stuttgart, the University of Rosenheim and several industry partners1.

INTRODUCTION Band saw blades are used especially for their small kerf and the possibility to process large logs in comparison to other timber cutting techniques. One of the disadvantages is the shorter endurance. The intention of this project is to enhance the durability of band saw blades. Present band saw blades have to be sharpened after approximately 8 hours. The project targets are:  enhance the durability until sharpening process ≥ 40 hours  enhance the fatigue strength ≥ 2.0 Mio. Alternations of loads (present 0.5 Mio.)  solidification of the Gullet  test on various stress levels The steps are following: 1. Simulation of the incurred forces and stresses, 2. Design and construction of an experimental test rig, 3. Simulation of the tooth force, 4. Fatigue strength experiments, 5. Analysis of the experimental results, 6. Realisation or a proto type band saw blade, 7. Industiral log break down experiments. 1

Tigra GmbH Oberndorf am Lech, Kähny Maschinenbau GmbH Backnang Germany, Alber GmbH & Co KG Ebersbach an der Fils, all Germany 71


SIMULATION OF FORCES AND STRESSES The simulation of the forces and stresses is based on different load cases. The calculation was carried out with the finite-element-analysis software ANSYS®. The results of this analysis are the stress distribution in the band saw blade and the identification of the area of maximum stress. The different load cases are a combination of pretensioning, bending stress and tooth force, with components parallel and perpendicular to the saw blade (cutting force and thrust force). The following table shows this load cases:

Figure 1: Example of load case #2 with horizontal and vertical forces The assumed parameters for the analysis is a bending stress level of 40.000 N and a tooth force of 150 N / tip. Table 1: Different load cases #

perpendicular

parallel

specifics

1

F1 =

150 N

F2 =

0N

load on every tooth

2

F1 =

150 N

F2 =

75 N

load on every tooth

3

F1 =

0N

F2 =

75N

load on every tooth

4

F1 =

150 N

F2 =

0N

load on one single tooth

5

F1 =

150 N

F2 =

0N

without saw blade prestress

6

F1 =

0N

F2 =

75 N

without saw blade prestress

7

F1 =

0N

F2 =

0N

only prestress

72


A

B

40.000 N 150 N

Figure 3: Tooth force

Figure 2: bending stress level

Figure 4: Stress level and maximum stress for load case 2 During a full cycle a single point in the gullet is subjected different loads and bending while running across the wheel. Figure 5 proves the bending stresses to be the dominant strain for the sawblade.

Figure 5: Maximum stresses of point A (see Fig. 6) during a full cycle.

73

20th International Wood Machining Seminar DESIGN OF A TEST RIG

A test rig was designed to carry out fatigue tests for different sawblades. It should be able to simulate a.) the bending stress b.) the tooth force a.) Simulation of the Bending Stress

2

Bending stress

Saw Blade Body Pitch Length = 25 mm

3

Tooth Height = 8.5 mm Gullet

1

Tensile stresses log

Blade Width = 22.0 mm

5 4 Teeth

6

Bending stress

Figure 6: Stresses in the band saw blade

Figure 7: band saw blade data

Base of the test rig is a standard joinery band saw (data: wheel diameter = 600 mm; band saw blade length = 4650 mm; speed 25m / sec.) allowing to downscale the dimensions of machine and blade by a factor 2 compared to a standard sawmill machine. This permits to reduce weight, space and costs. Also the size of band saw blades can be reduced from 10 m length (as normally used in saw mills) down to 4 m. The width is reduced from 200 mm to 35 mm which makes experiment and analysis easier.

b.) Simulation of the Tooth Force

Figure 8: Gear for simulation of tooth force

Figure 9: Gear in action

The tensile stresses caused by the prestressing force, is measured by two load cells on the upper wheel, while the bending stress may simply be calculated from blade thickness and wheel 74

20th International Machining Seminar diameter. The number of revolutions of the Wood wheels is monitored by a photo sensor in order to determine the exact number of load cycles until failure.

The tooth force during log break down is simulated by a gear, strained by a eddy-current brake.

FATIGUE STRENGTH EXPERIMENTS Table 2: List of different band saw blades tested #

width

Tensile stress

tooth gullet

fatigue cycles

complete crack

V1 V2 1

[mm] 0,6 0,6 1

[N/mm²] 180 180 280

2.206.312 588.201 120.941

no yes yes

2

1

280

normal (H.H.)

107.942

yes

83

69

3

1

280

normal

77.250

yes

44

18

4

1

280

normal (H.H.)

113.127

yes

207

28

5

0,6

280

normal

218.470

yes

6

2

6 7 8

not occupied 0,6 250 0,8 250

normal normal

999.993 240.909

no yes

3 18

0 8

normal normal normal

no of cracks ≤1,0mm >1,0mm 0 0 0 1 141 22

9

1

250

normal

224.215

yes

1

3

10

0,8

250

polished & chamfered edge

336.844

yes

6

0

11

0,8

250


999.993

no

0

0

12

1

250


1.087.474

no

0

0

13

1

250


1.208.305

no

0

0

14

1

250

Sandvik Durashift

407.254

yes

5

10

15

1

250

Sandvik Durashift

259.152

yes

8

26

16

0,8

250

Sandvik Durashift

615.215

yes

3

2

17

0,8

250

Sandvik Durashift

786.896

yes

2

3

18 19

free 1,0

250

chamfered edge, rolled

170.986

yes

1

23

20

1,0

250

chamfered edge, rolled

172.762

yes

0

8

21

0,8

250

normal

450.056

yes

1

4

22 23

0,8 0,8

250 250

normal Thermex

1.091.036 351.819

no yes

0 7

0 44

24

0,8

250

Thermex

555.499

yes

58

17

25

1,0

250

Thermex

281.954

yes

17

3

26

1,0

250

Thermex

133.353

yes

26

39

27

1,0

250

2-step grinding

107.134

yes

24

133

28

1,0

250

2-step grinding

188.831

yes

17

137

29

1,2

250


216.623

yes

0

2

30

1,2

250


142.378

yes

0

2

In order to enhance the fatigue strength of the saw blades following methods were applied:  Chamfering of the edges  Polishing of the gullet  Compaction by rolling and peening 75

 Thermal treatment 20th International Wood Machining Seminar  Enhanced strength of blade material (Durashift) The strain of the saw blades was varied by the thickness and the pretensioning level. All together 30 band saw blades were tested. The following table shows the different band saw blades ad the achieved results. The fatigue tests of a band saw blade could have two different results:  Failure occurs before 2.0 Mio cycles are reached  2.0 Mio. cycles of loads are reached without failure. These blades are assumed to have a fatigue strength beyond the load applied. After the tests the band saw blades are scanned for additional cracks in the gullet. The number of cracks were determined and monitored. Every identified crack is a failure source having started on a weak spot within the gullet. In general these initial cracks are generated by the standard sharpening process since the abrasive wheel is moving perpendicular to the gullet. In total 29 blades treated with different methods to improve the fatigue strength were tested. An overview is given in table 2.

ANANLYSIS OF THE EXPERIMENTAL RESULTS The results are shown in figure 12. It shows clearly that among all variations only the blades with polished gullets reached a fatigue strength beyond 600 N/mm². All other measures achieved only moderate increase of the lifetimes, far from the required target of 2.0 Mio cyles.

0.2 mm

8.1 mm

Figure 10: Crack (size < 1,0 mm)

Figure 11: Crack (size > 1,0 mm)

so

run through

0,6 mm R/7

700 N/mm²

0,8 mm R/7 1,0 mm R/7 0,8 mm polished

600

1,0 mm polished

N/ mm²

1,2 mm polished 0,8 mm Durashift 1,0 mm Durashift

500

N/ mm²

1,0 mm gerollt 0,8 mm Thermex 1,0 mm Thermex 1,0 mm fine

400 N/mm²

ground

300 N/mm² 1x 106

1x 109

2x 109

Figure 12: Results of fatigue stress experiments 76

4x 109

alternation of loads

Microscope Analysis


Microscope analysis size = 0.5x0.5mm² top view

Grinding burr

Left Saw blade surface ploished Right Saw blade surface standard

Figure 13: Microscope results of a polished surface and a normal surface

The gullet surface of different band saw blades and different cracks were analysed and measured by microscope and a chromatic roughness gauge. Figure 13 displays examples. The grinding burr caused by a standard wheel running perpendicular with the saw blade can be seen very clearly. This burr is one of the main reasons for triggering fatigue cracks. This effect could only be avoided by a different grinding direction or polishing with small grinding bits. Also a two step grinding process with finer finish but same grinding direction did not accomplish the expected results.

REALISATION OF A PROTOTYPE SAWBLADE In order to verify the results of the experiments and analyses a couple of full sized prototype band saw blade were produced. First step was to calculate the stress distribution in this blade. This stress distribution was analysed in the same way like the other band saw blades. The saw blade was manufactured by the industrial partners and tested in the saw mill of the University of Rosenheim. Steps to create the proto type band saw blade were:  Calculation of maximum stress level  Brazing of carbide tips  Polishing the gullet  Break down experiments in the saw mill

Figure 14: Prototype band saw blade - Calculation of maximum stress level.

Figure 15: Prototype band saw blade

77

International Machining Seminar The breaks down experiments20th at the UniversityWood Rosenheim were successful. No carbide tip was lost though the test time was limited due to the restrictions of a laboratory not being designed to process large quantities of material. Also the achieved quality of the sawn surfaces was acceptable. However first experiments under industrial conditions suggest that further reinforcement of the tip-saw-blade interface will be required.

CONCLUSION Important prerequisites for the production of carbide tipped band saw blades were developed. In particular the fatigue crack problem was resolved and the stability of the interface for carbide tipping improved. By polishing and chamfering the gullet a fatigue strength for standard saw blades beyond 600 N/mm² was achieved. The other techniques tested achieved only minor improvements not sufficient to Figure 16: Breakdown experiments at suppress fatigue cracks completely. The interface for the University Rosenheim carbide tipping, an enhanced brazing technique attained sufficient stability for Laboratory conditions. First tests under industrial conditions indicate however that further improvement will be required.

List of contents List of figures ............................................................................................................................ 8 List of tables .............................................................................................................................. 9 Introduction ............................................................................................................................... 1 Simulation of forces and stresses .............................................................................................. 2 Design and construction of a test rig ......................................................................................... 3 a.) Simulation of the bending stress ...................................................................................... 4 b.) Simulation of the tooth force ............................................................................................ 4 Fatigue strength experiments .................................................................................................... 5 Analysis of the experimental results ......................................................................................... 5 Microscope analysis .............................................................................................................. 7 Realisation of a proto type band saw blade ............................................................................... 7

List of figures Figure 1: Example of load case #2 with horizontal and vertical forces ........................................ 2 Figure 2: bending stress level ....................................................................................................... 3 Figure 3: Tooth force .................................................................................................................... 3 Figure 4: Stresses in the band saw blade ...................................................................................... 4 Figure 5: band saw blade data ....................................................................................................... 4 Figure 6: Gear for simulation of tooth force ................................................................................. 4 78

20th International Wood Machining Seminar Figure 7: Gear in action ................................................................................................................ 4

Figure 8: Crack (size < 1,0 mm) ................................................................................................... 6 Figure 9: Crack (size > 1,0 mm) ................................................................................................... 6 Figure 10: Results of fatigue stress experiments .......................................................................... 6 Figure 11: Microscope results of a polished surface and a normal surface .................................. 7 Figure 12: Protoy type band saw blade - Calculation of maximum stress level. .......................... 7 Figure 13: Proto type band saw blade ........................................................................................... 7 Figure 14: Breakdown experiments at the University Rosenheim ............................................... 8

List of tables Table 1: Different load cases ........................................................................................................ 2 Table 2: List of different band saw blades tested ......................................................................... 5

79


Optimization of the choice of the cutting materials for solid wood tools P-J. MÉAUSOONE, S. AUCHET LERMAB (Laboratoire d’Etudes et Recherche sur le Matériau Bois) 27 rue Philippe SÉGUIN, B.P. 1041 – 88051 Epinal CEDEX 9

ABSTRACT The choice of a tool is generally made on geometric notions, sometimes on angular values, search for an important lifetime of the tool but without worrying about the precise quality of the cutting material. Indeed, the wear is a phenomenon badly managed within the wood industry. Wood as heterogeneous and anisotropic material, have a structural complexity making difficult the establishment of general laws of behaviour. Furthermore, companies are SME and SMI which could not easily made research and development. The relation during cutting process between machine, tool and wood pieces takes us multi parameters to understand and study. The aim of the project is to understand the different effects of a choice and to propose various courses to optimize this choice of woodworking tools. In a first time, the research carried on wear of tool: analysis and understanding of the impact of the corrosion and the abrasion on the wear of a cutting edge, effect of the cutting temperature on the evolution of the wear, evaluation of wear resistance for different cutting materials in industrial test. After, we have utilized selected tools on machines with different technical approach: tool-spindle attachment, connection of teeth on the tool; measurement of effort and power during wood process. We propose to present some results of the combination of abrasion, corrosion and temperature action on cutting materials, with application on woodworking tools in industrial utilization.

1. MECHANISMS OF TOOLS’ WEAR IN WOOD PROCESSING Wood processing by removal chips can be characterized by the contact between a tool and the material to be manufactured. This one engenders important frictions between the active surfaces of the tool, the surface finished of the product and the chips. We can determine also mutual actions giving stress of flexion, compression or shear. These phenomena are factors occurring in the degradation of the cutting edge of the tool. Wood processing is in this logic and for a long time, the industry and the research team work on the hardness and the impact strength of materials for cutting. However, wood is a natural composite with heterogeneous structure and a complex chemical structure consisted of carbon, oxygen and hydrogen. So, we wished to integrate the third dimension into this concept, which is the corrosion resistance (figure 1). Indeed, the search of the perfect material does not have to be any more a hard and flexible material, but a hard, flexible and resistant material in the corrosion.

Fig n°11: Evolution of characterization of materials for wood manufacturing

80


Generally, we suppose that the mechanical and chemical phenomena are the main causes, but we forget the notion of temperature which increases the wear of the tool. It determines the fourth dimension to the analysis which we wish realized. In wood industry, many users think that, the more the cutting material is hard, the more the tool possesses a long life cycle. The quantification of the life cycle remains evasive, because depending on parameters resulting from a complex material, wood, and bad conditions of manufacturing.

2. OUR APPROACH OF THE WEAR PHENOMENA We considered that the wear of a tool was due to several phenomena in interaction [1]. Indeed, as described in the figure 2, we think that the wear is inferred by two main mechanisms: first of all, attacks constituted by a corrosive attack decreasing the mechanical resistance of the surface, then by an abrasive attack, facilitated by the decline of the resistance of the surface. This first cycle is for us prevailing within the framework of wear, giving a recession of the cutting edge, with a linear increase. Naturally, then, the repeated or occasional shocks take place in the process of degradation of the cutting edge. If a shock which exceeds the mechanical resistance of the cutting material, a part of the cutting edge will come away. But a shock can create a breach on the surface and so accelerate the process of corrosion which generates an acceleration of the wear by abrasion. Other factors, such as temperature or conditions of manufacturing intervene within these various processes, with a strong increase of the wear. Indeed, the importance of attacks are determined by the range of temperature in which they are, according to the conditions of manufacturing (for example, too important thicknesses of chips or on the contrary, too low. Abrasion Wear cycle 1

Corrosion

Shocks Wear cycle 2

Fig n°2: Tools wear cycles

3. MUTUAL ACTION OF ABRASION-CORROSION Initially, we approached the research about wear on the corrosion phenomenon due to the wood [ 2 ] . The observation of the damages by corrosion let think that the mechanical properties of the cutting face of tool exposed to the corrosion are decreased and that, therefore, the damages due to the abrasion or to the shocks are facilitated. So, we are certain that the corrosion plays a role in the process of degradation of a cutting edge, but we cannot quantify the importance of this phenomenon yet (figure n°3 and 4).

Fig n°3: Example of SEM (Scanning Electron Microscopy) image of the corroded zone and optical section of steel

81


Fig n°4: Example of SEM (Scanning Electron Microscopy) image of the corroded zone and optical section of carbide The research about degradation by abrasion was decomposed into two parts: a general study of the manufacturing wood to conceive a test bench (TEEMO: Test of Studies and Damages on Tools’ Materials) dedicated to the study of the abrasion in the conditions near as possible to the reality and then a series of tests made thanks to the bench (figure n°5). During preliminary tests made on a pawn disk, when we made a certain distance of friction, a deposit of the hardest material formed in the contact area. It is naturally impossible to qualify the mechanisms of wear if a deposit intervenes between the tool and the studied material. As we wished to manage at best the wooden quantity to make the tests, we arrived at the conclusion that it was necessary to regenerate the surface rubbed just after friction. We thus looked for a means to clean this surface during the test. The reserved solution is the one which consists in putting in series two vertical spindle with each a tool: the one realizing the operation of friction and the other one really manufacturing the board.

Refresh spindle shaper

Test tool 1

2

4 3

Fig n°5: TEEMO and cycles of cutting 1: Cutting phase of board 2, 3, 4: Phases for initial position of the wooden board with a rectangular cycle.

82


Finally, to show the interaction between corrosion and abrasion, we realized test tubes "treated" pieces, and then tested on the bench TEEMO. We were able to notice a difference of behavior enter corroded and not corroded steel [ 3 ]. We notice that the modes of degradation are not simply stressed by the corrosion but change nature. Indeed, without corrosion, the process of degradation takes place rather by separation of layers, but, with a phase of corrosion before the friction, the damages are more due to a decomposition of the matrix with a wrench of grains. So, interaction of the physical-chemical properties as "hardness / corrosion resistance" is more significant against the abrasive wear of a tool, than the only factor of hardness. As regards carbide, the phenomenon of wear is different from steel. A not corroded carbide is going to resist the friction with a light and regular wear; but this analysis is to be balanced because the surface of friction seems to be too important for this type of material. Nevertheless, carbide with a small quantity of binder resists less the micro-shocks engendered by the mode of circular "cutting" used in wood industry. On the other hand, the corrosion takes a more degradation, as we suppose in our initial hypotheses. Now the difficulty of this kind of composition comes from the subtlety of the dosage between grains of carbides and binder.

4. TEMPERATURE EFFECT Among all the factors which understand the wear of a tool, temperature influences on practically all the others. In our concept in three dimensions (tenacity, hardness and corrosion resistance) the temperature is the fourth dimension. If a material has the hardness the most adapted with the good tenacity and the good corrosion resistance but if these three properties are obtained in a temperature different from cutting situation, the material will become, actually, potentially inappropriate. The research on tool-material interaction for metallic materials included the interest of this datum and developed techniques which allow knowing the temperatures generated by manufacturing steel. As regards the research wood, the interest asserted itself only recently. It considered for a long time that the temperatures during the cutting were unimportant, given that the wood after manufacturing was not "burned". 4.1 PROBLEM OF MEASUREMENT Nevertheless, the wood carbonizes between 280°C and 300°C [4] and a first approach would determine a range of possible value for the cutting temperature. The anatomical structure of wood is a good insulator heat, known and used by all for this specificity. So, the flows of heat generated by the cutting are evacuated by the cutting edge and not in the material wood, so accumulating the heat in the tool rather than in the wood. Furthermore, if a board is stopped during a manufacturing (even some µs), we notice invariably a track of burn on the wood after manufacturing. By taking into account all the data quoted previously, we can think that the temperature of the tool rises at least until 280°C. Darmavan [5] measured, during a study on the wear of tools for manufacturing of panels (like wood cement), temperatures reaching 350°C for Douglas species, with a cutting speed of 60 m/s. The method used to measure the temperature is diverted from the method of the thermocouple. On this principle, it created a couple of materials establishing the cutting edge. It is interesting to note that in this study, the density of the material cannot explain the rise of temperature. Indeed, materials of lower density such as the Douglas species (0.32), generate temperatures higher than the "wood cement of low density" (0.81).

83


4.2 TRADITIONAL APPROACH The first choice concerned to a tool with two teeth to realize these tests. The equipment is established by a body of traditional tool with a 60 mm boring to receive a hydraulic muff mounting sleeve assuring the centring of the tool on the spindle. On this cutter spindle, two specific teeth are taken to obtain cutting diameter of 150 mm and a cutting height of 50 mm. To cover this tool, a support consisted of a rotating collector and two thermometers with resistance in each tooth was realized (figure n°6).

Fig n°6: temperature measurement by rotating collector We don’t present the different results, obtained with this system. Indeed, this manipulation is not a new approach, but we utilized the results in comparison with the second approach of temperature measurement. 4.3 INDIRECT MEASUREMENT

Dureté HV

Research based on the knowledge of the cutting temperature is not in spare number for wood processing. But, new methodologies were elaborated to answer this problem. So, Kusiak [6] used the inverse method to determine the temperature in the case of MDF lathing. The indirect or inverse measures consisted in estimating the temperature of the cutting zone from measures made for distances different from the targeted zone. To know more exactly the temperature of the cutting edge, the principle which we wished developed was based on tests of measure of the temperature based on the changes of mechanical property of steel (figure n°8). So, we measured the loss of hardness for a hardened steel during a reheating, and determine in an equivalent way the reaction of a cutting edge during manufacturing. This second phase will allow us to link the result to the temperature measured by the first thermometer resistance in the reality. So, we can build a function of transfer between the reality and the temperature found by the first tests. 5 9 05 9 0 5 7 05 7 0 5 5 05 5 0 5 3 05 3 0 5 1 05 1 0 4 9 04 9 0 4 7 04 7 0

0

0

50

50

100

100

1 50

200

150

2c o0u rb0e d e re v e n u

250

250

300

300

C o u rb e d e re v e n u

Figure n°8: Example of tempering

84

350

400

400 350 T e m p é ra tu re e n °C


The method which we used appears in the following way: Stage n°1: first of all, we realized the tempering kinetics of the 35NCD16, the reference steel used for the test to establish the relation enter the definite time in min, the temperature in °C and the Vickers hardness (Figure n°9).

Dureté

Fig n°9: Tempering kinetics of 35NCD16 Stage n°2: on a cutting edge, subjected to a manufacturing on oak during 15 min, we realized the measure of micro hardness to be able to know about temperature on cutting edge (figure n°10)

Fig n°10: Micro hardness of teeth after manufacturing

85


Stage n°3: knowing the hardness of both metals, we made correspondence of the various values (Tab n°2): Speed rate in Rds/min

Feed rate (m/min)

Stroke (mm)

Average thicknees for 80 µm

Temperature °C

Average thicknees 200 µm

Temperature °C

3000

5

1

607

120-130

612

110-120

8000

15

1

593

140-240

594

140-240

5500

25

1

583

230-240

594

140-240

3000

25

3

614

110-120

610

110-120

8000

5

3

559

310-320

570

280-300

5500

15

3

616

110-120

600

120-130

Tab n°2: Cutting temperature according to cutting conditions. Indirect measure by modification of the material hardness.

5. SHOCKS The notion of shocks can be described by several manners: indeed, the teeth of tools go in and out of the material; can meet disruptive elements of the material to be manufactured; they are directly bound to the stability of the tool and, finally, to its mode of fixation on the tool. Contrary to other a manufactured material, wood is very heterogeneous. Very important differences of density thus can be present in the same board of wood. These differences of density are stressed in softwood with presences of knots. The density of the knot can reach two-three times the density of the wood. For example, the circular saw blades undergo during the sawing of the shocks with particles of foreign bodies (metal, stones) very hard which damage carbide tips [7]. The split or missing teeth are factor for out of balance during rotation. Furthermore, the breaking of two or three consecutive tips is more harmful for the dynamic balance than the breaking of twice two tips diametrically opposite … The pitch of tooth, is doubled or tripled because of the break teeth, so that the tooth placed after the damaged zone takes a chip thickness two or three times as important as the normal, provoking stripes on the sawed surface. This more important chip thickness can also pull the deterioration of this tooth, provoking an effect "in chain" of teeth breaking (figure n°11).

Fig n°11: Broken teeth on circular saw

86


6. CONCLUSION The knowledge of the wear in the manufacturing process is determining, with the wish to anticipate the rulings of production and the products quality. Research for the aggravating factors such as the corrosion or the abrasion allowed us to include better the phenomenon of degradation of the cutting edge. The temperature intervenes as accelerator of these factors. Our original method allows us to approach temperatures supposed during the manufacturing of the wood. It has however the handicap to be destructive. The highest temperatures are in correlation with the highest stroke. The various essays allow us to look for steels used as tool steel in the wood sector can reach their peak of performance around 300-350°C (minimal value for the cutting edge). But, at this moment, there are only reports for shocks; but, later, it will be necessary to determine the real part of the shocks in the notion of tool wear in wood manufacturing. REFERENCES: 1. Gauvent M., Optimisation de la durée de vie d’un outil de coupe pour l’industrie du bois. Analyse et compréhension des modes d’usure. Mise au point de solutions innovantes avec tests industriels, University of Nancy PhD Thesis, with partnership of CM2T company – Nancy (France), 2006 2. Gauvent M, Rocca E., Méausoone P-J., Brenot P., 2006 - Corrosion of materials used as cutting tools of wood. WEAR Journal, Volume 261, Issue 9, pp 1051-1055. 3. Gauvent M, Méausoone P-J., Martin P., Rocca E.,2005, ”Impact of abrasion and corrosion by solid wood in tool's wear”, 17th International Wood Machining Seminar, Rosenheim, pp. 571-280, septembre 2005. 4. Bernard C, Caractérisation et optimisation de bois fragmenté en chaufferie automatique, University of Nancy PhD Thesis – Nancy (France), 2005 5. Darmavan W., Tanaka C. , Usuki H. , Othtani T. , Performance of coated carbide tools in turning wood-based matérials: Effect of cutting speeds and coating materials on the wear characteristics of coated carbide tools in turning wood-chip cement board, Journal of Wood Science, vol. 47 pp 342-349, 2001 6. Kusiak A., Caractérisation thermique des outils revêtus en usinage du bois. Thèse en Sciences du Bois, Université de Bordeaux1, 172p, 2004 7. Simonin G. – Amélioration de la performance des outils pour la première transformation du bois - University of Nancy PhD Thesis, with partnership of SIAT BRAUN company – Urmatt (France), 2010

87


Wear Characteristics of Newly K10 Coated Cutting Tools in Cutting Particleboard Wayan DARMAWAN(1), Dodi NANDIKA(1), Hiroshi USUKI(2), Rémy MARCHAL(3) (1) Bogor Agricultural University, INDONESIA / [email protected] (2) Shimane University, Matsue, JAPAN / [email protected] (3) ParisTech, Cluny, FRANCE / remy.marchal @cluny.ensam.fr

ABSTRACT This paper presents delamination wear characteristics on clearance face of newly K10 coated cutting tools when milling particleboard. The newly coated K10 cutting tools were monolayer TiAlN, and multilayer TiAlN/TiSiN, TiAlN/TiBN, and TiAlN/CrAlN. Particle board of density of 0.61 g/cm3 was cut by using these coated cutting tools. Cutting tests were performed at a high speed cutting of 41.8 m/s and feed rate of 0.2 mm/rev to investigate the delamination wear characteristics on the clearance face of these coated tools. Experimental results showed that the newly multilayer coated tools (TiAlN/TiBN, TiAlN/TiSiN, TiAlN/CrAlN) retain smaller amount of delamination wear than the monolayer coated tool (TiAlN) in milling the particle board. The best multilayer coated tool was the TiAlN/AlCrN which only suffers a slight chipping on the cutting edge at the final cutting length. High hardness, low coefficient of friction, high resistance to oxidation, and high resistance to delamination wear of the multilayered TiAlN/CrAlN indicate a very promising applicability of this coating for high speed cutting of abrasive wood based materials. Keywords : particleboard, newly multilayer coated tools, delamination wear, high speed cutting, oxidation

INTRODUCTION Particleboard has been made in large and increasing quantities in many countries. Recently, the use of the particleboard has been increasing for building constructions and decorative purposes. In the secondary wood manufacturing industry, where wood-based material such as particleboard is machined extensively, tool wear becomes an important economic parameter. Therefore, investigating the wear characteristics will lead to making better choices of cutting tool materials used to cut wood-based material. Machining of particleboards causes cutting tools to wear out much faster than machining of solid woods. Rapid dulling of steel cutting edge of router bit, saw teeth, or other knives blade in machining the particleboards has been a well-known phenomenon. The use of tungsten carbide tools, which are now widely used in the wood working industry, especially for wood-based materials are also limited because of relatively high rate of wear. This fact was approved by some findings that the tungsten carbides were susceptible to wearing during cutting particleboard and fiberboard due to high temperature oxidation and abrasion [1,2]. Thus an effort of coating the surfaces of the carbide cutting tool with a hard coating material has been already developed in order to increase the wear resistance of the carbide tool.

88


Some coatings, that have been commercially produced, consist of a monolayer of titanium carbide (TiC), titanium nitride (TiN), titanium aluminum nitride (TiAlN), titanium carbonitride (TiCN), chromium nitride (CrN), chromium carbide (CrC), and diamond-like carbon (DLC). These coatings have been commercially used in the metal working industry with a good result [3]. However their applications for wood-based material machining, which is realized under high speed cutting, are limited because of the occurrence of coating film delamination. K10 tungsten carbide inserts were coated with TiN and TiC films by chemical vapor deposition (CVD) method for cutting particleboard [4]. It was noted in the work that TiN coating brought no advantages in the milling particleboard, i.e. the edge life attained or the total tool paths, respectively were not longer than those attained by uncoated cutting edges. TiC coated tungsten carbide tool was also reported to show in large edge fractures after a short time of operation. K grades of tungsten carbides were coated with TiN, Ti(N, CN) and TiAlN2 by physical vapor deposition (PVD) method for continuous milling of particleboard [5]. It was noted in the study that a slight improvement in wear resistance was provided by coatings which were synthesized to carbide grade with fine grain size (0.8 µm) and low cobalt content (3%). In contrast, coatings applied to carbide grade with higher cobalt content and coarse grain size were reported to decrease the wear resistance of the tools. The primature failure mode for the PVD coatings tested in the work was chipping of coatings on the rake face. This fact was caused by inadequate adhesion of coating over the tool surfaces due to improper substrate and coating combination. TiN, TiAlON and TiC coatings were synthesized on the surface of tungsten carbide (93.5% WC, 5% Co, 1.5% (Ta, Nb)C by plasma assisted chemical vapor deposition (PACVD) for milling laminated particleboard [6]. It was noted in the work that TiN and TiAlON coated carbide tools did not provide any improvement in wear resistance compared to uncoated carbide tool, and the TiC coated carbide tool provided only a slight improvement. Further, it was reported in another study that the wear of TiN, CrN, CrC, TiCN, TiAlN coated in cutting wood-chip cement board occurred due to delamination of coating film in both low and high speed cuttings [7]. The TiAlN was found to retain the lowest occurrence in delamination among them. Delamination of these coating films for high speed cutting was caused by oxidation, which was accelerated by an increase in cutting temperature. The above references indicate that the monolayer coatings did not provide significant improvement in the cutting tool life for high speed cutting of wood based materials. Therefore ongoing research efforts should be proposed to achieve better performance of the coated carbide tools in cutting wood-based materials. Multilayer coatings would be a promising technique to improve the performance of the monolayer coatings. TiAlN coating, which is high in hardness, good in oxidation resistance and better in wear resistance among the other monolayer coatings, was multilayered with the newest generation coatings of titatnium boron nitride (TiBN), titanium silicon nitride (TiSiN), and chromium aluminum nitride (CrAlN) which have been noted to keep excellent properties (high hardness, low friction coefficient, high oxidation and corrosion resistances) [8,9]. In this work, TiAlN coating was multilayered with TiSiN, TiBN, and CrAlN coatings onto the surface of K10 tungsten carbide using PVD method. The newly multilayered coatings tools (TiAlN/TiBN, TiAlN/TiSiN, TiAlN/CrAlN) were experimentally investigated for the possibility of their application in machining particle board. The purpose is to investigate the delamination wear characteristics of the newly coated tools in high speed cutting of particle board.

89


MATERIALS AND METHOD Multilayered-Coated Cutting Tools and Work Materials General specifications of the newly coated cutting tools tested, and particleboard machined are shown in Table 1 and Table 2 respectively. K10 carbide tool (94% WC, 6% Co) that was selected as a substrate was 16 mm long, 16 mm wide and 6.2 mm thick. The hardness of the K10 was measured to be 1400 HV. A 65o wedge angle used in this experiment is now being commercially produced especially for cutting wood-based materials. The K10 carbides were coated with monolayer coating of TiAlN and multilayer coatings of TiAlN/TiBN, TiAlN/TiSiN, TiAlN/CrAlN by PVD method on both rake and clearance faces. Table 1. Specifications of coated carbide tools tested Coating material TiAlN TiAlN/TiBN TiAlN/TiSiN TiAlN/CrAlN

Film thickness (μm) 3 3 3 3

Hardness (GPa) 28 44 35 38

Oxidation temperature start at 600oC start at 700oC start at 700oC start at 800oC

Friction coefficient 0.91 0.56 0.61 0.45

Table 2. Specifications of the particle board machined. Characteristics

Value

Thickness

12 mm

Moisture Content

8%

Density: Average Outer part Inner part

0.61 (g/cm3 ) 0.73 (g/cm3 ) 0.51 (g/cm3 )

Experimental Setup Cutting test was set up on the numerical controlled (NC) router, as shown in Figure 1. The particleboards were prepared in form of rectangular (12 mm x 1000 mm x 1500 mm). A piece of particleboard was placed on the table of the NC router and locked by some screws. A cutting tool edge was held rigidly in a tool holder of 32 mm cutting circle diameter. The condition of cutting is shown in Table 3. Cutting was performed a long the edge of the board by setting spindle rotation in clockwise direction. The movement of the board during cutting was controlled by feeding directions of the table with the order of 1 and 2 continuously as shown in Figure 1. The first feeding caused the board to be edged in down milling direction and the second feeding caused the board to be edged in up milling direction. To understand the effect of density along the thickness of the board, 6 mm (3 mm outer part and 3 mm inner part) depth of cut was performed. Density of inner part of the board was 0.51 g/cm3, and of outer part of the board was 0.73 g/cm3.

90


Table 3. Milling conditions Variable

Condition

Cutting speed Feed Spindle speed Feed speed Width of cut Depth of cut Diameter of cutting circle Rake angle

41.8 m/s 0.2 mm/rev 25000 rpm 4000 mm/min 2 mm 6 mm 32 mm 8o

Clearance angle

7o

Tool holder Particle board

Figure 1. Photograph of milling on the edge of the particle board

Measurements The coated tested tools were inspected before testing to assure that there are no surface cracks and defects of coating film on both the rake and clearance faces using an optical video microscope. The cutting was stopped at every specified length of cut (100 m), then delamination of coating film were measured along clearance faces of the tools. Measurements of delamination wear on the clearance face were made by using an optical video microscope as shown in Figure 2. The tools were also inspected at the final cut using a scanning electron microscope/energy dispersive spectroscopy (SEM/EDS) for identification of the mode of cutting edge failure and oxidation occurrence.

91


Initial edge Substrate Edge recession

Coating film

Delamination wear

65o

Figure 2. Schematic diagram of delamination wear measurement on the clearance face of cutting tool

RESULTS AND DISCUSSION Delamination wear behaviors on clearance faces of the newly coated tools are provided in Figure 3. The results in Figure 3 indicate that the amount of delamination wear increased with increasing in cutting length. The multilayer coated tools provided better performance especially in reducing the progression of delamination wear than the monolayer TiAlN in cutting the particle board. Though the monolayer TiAlN and multilayer coated tools showed almost the same delamination progress near beginning of cutting, however the delamination wear of the monolayer TiAlN increased markedly and exceeded the delamination wear of the multilayer coated tools, which retained gradual delamination progresses during cutting the particle board. The monolayer TiAlN cutting tool suffered delamination wear of about 122 µm for cutting outer part of the particle board and about 60 µm for inner part of the particle board at the 1 km cutting length. Otherwise, the delamination wears of the multilayer coated tools were less than 100 µm for the outer part of the particle board and less than 50 µm for the inner part of the particle board at the 1 km cutting length. Further, the highest rate of delamination wear was retained by the monolayer TiAlN followed by multilayer TiAlN/TiBN and TiAlN/TiSiN. The lower in hardness, lower in oxidation resistance, and higher in friction coefficient of the monolayer TiAlN compared to the multilayer coated tools (Table 1) would be the reason for this phenomenon. It also appears that the outer part of the board (Figure 3b) caused higher amount of delamination wear compared to inner part of the board (Figure 3a). The outer part of the particle board caused delamination of TiAlN, TiAlN/TiBN, and TiAlN/TiSiN coatings in the rate of about 1.5X faster than the inner part of the board. The structure of the inner part of the board was more porous than that of the outer part. Therefore, more materials are machined at a given cutting volume for outer part of the board. The higher adhesive content in the outer part imposed higher abrasion during the cutting. However, no delamination wear was observed for the multilayer TiAlN/CrAlN both in cutting the outer and inner parts of the particle board (Figure 4-b4).

92


(b)

(a)

Figure 3. Delamination wear behaviors of the newly coated tools with cutting length in milling particle board

The results in Figure 3 and Figure 4 indicate that the multilayer TiAlN/CrAlN has the highest resistance in delamination wear compared to the multilayer TiAlN/TiSiN and TiAlN/TiBN in cutting the particle board. The TiAlN/CrAlN did not suffer any delamination of coating film up to cutting length of 1000 m. The cutting edge of the TiAlN/CrAlN suffered slight chipping of coating film at the cutting length of 800 m, and still retained the slight chipping up to the final cutting length without followed by delamination of coating film (Figure 4-b4). The high delamination wear resistance of the TiAlN/CrAlN is considered to be due to the two following reasons. First, friction coefficient of the multilayered TiAlN/CrAlN coatings was lower than that of TiAlN, TiAlN/TiBN, and TiAlN/TiSiN, which will lead to exert less abrasion against hard abrasive materials contained in the particle board. The hard abrasive materials in the particle board consisted mainly of cured adhesive, which generates harmful effect on the edge of the coated cutting tool. Second, oxidation temperature data in Table 1 suggest a better oxidation resistance retained by the multilayered TiAlN/CrAlN coatings compared to the TiAlN, TiAlN/TiBN, and TiAlN/TiSiN. This phenomenon could be confirmed by previous result in which delamination of monolayer coatings (TiN, TiAlN, TiCN, CrN) in high speed cutting of wood chip cement board is caused by higher contribution of oxidation [10]. The oxidation was reported to be accelerated by the increase in cutting temperature up to 800o C due to increasing in cutting speed above 30 m/s. A possible high temperature generated during high speed cutting of particle board in this experiment would oxidize the TiAlN, TiAlN/TiBN, TiAlN/TiSiN coated carbide tools, which lead to the severe delamination of coating films due to thermal degradation. Investigations of worn edges of the coated tools under an optical video microscope show that relatively the same delamination mechanisms of coating films were observed in cutting the particle board. The results in Figure 5 depict the mechanism of delamination of TiAlN/TiSiN coatings which is selected for discussion in this article. Delamination of the TiAlN/TiSiN coatings was proceeded by slight chippings of coating film at the cutting edge. The extent of chipping was investigated to increase at the 200 m of cutting length. As the cutting was continued up to cutting length of 300 m, the cutting edge underwent more prominent chippings of coating film. Chipping of coating film occurred on the whole cutting edge as the cutting length reached 400 m. Further, the TiAlN/TiSiN coating films were gradually delaminated in proportion along the cutting edge as the cutting action was continued above 400 m. It is

93


considered that the wear of the K10 substrate occurred as the TiAlN/TiSiN films were disappeared from the substrate.

K10 substrate

Cutting edge

a1 Delamination line Clearance face

TiAlN

a2

TiAlN/TiBN

a3

TiAlN/TiSiN

a4

TiAlN/CrAlN

b1

Coating film

b2

b3

Chippinf of coating film

b4

Figure 4. Wear patterns of the coated cutting tools before and after 1 km cutting length SEM micrographs of the worn edges of the TiAlN, TiAlN/TiSiN, TiAlN/TiBiN, and TiAlN/CrAlN coated tools were presented in cutting the particle board for 1 km cutting length (Figure 6). The SEM micrographs reveal that relatively the same patterns of edge wear were generated by TiAlN, TiAlN/TiSiN and TiAlN/TiBiN. Though severe delaminations were generated by the TiAlN, TiAlN/TiSiN, and TiAlN/TiBiN coated carbide tools, however delamination did not occur in the TiAlN/CrAlN coated carbide tool. This fact will suggest that the substrate of the TiAlN, TiAlN/TiSiN, and TiAlN/TiBiN coated tools would be exposed to any possible mechanical abrasion, which caused retraction of tungsten carbide grains from the substrate during the cutting (Figure 6a-c). Retraction of carbide grains was reported to cause

94


corrugated cutting edge which tends to produce rough quality of board surfaces during the cutting [10]. Otherwise, TiAlN/CrAlN coatings strongly covered and protected the K10 carbide substrate from wearing, and maintained sharp cutting edge (Figure 6d). Initial edge

500 m 0m 600 m 100 m

primature 800 m

200 m

severe chipping

300 m 1000 m delamination

400 m

Figure 5. Wear mechanism of the TiAlN/TiSiN multilayered coating in cutting the particle board

Cutting edge

1000x

TiAlN

K10 substrate

a

1000x

TiAlN/TiBN

b

Chipping

1000x

TiAlN/TiSiN

c

1000x

TiAlN/CrAlN

d

Figure 6. SEM micrograph of worn edges of the newly coated cutting tools in cutting particle board at the final cutting length

95


CONCLUSIONS It could be pointed out the following conclusions based on the findings of this study. 1. TiAlN/CrAlN multilayered coatings are superior in reducing the progression of delamination wear in cutting the particle board. 2. Delamination of coating film for TiAlN, TiAlN/TiSiN, and TiAlN/TiBiN is proceeded by the occurrence of chipping of coating film mainly caused by mechanical abrasion. 3. Wear pattern of the coated carbide tools in cutting the particle board is almost the same, in which the wear of the carbide substrate occurs after the coating film was disappeared from the carbide substrate. 4. High hardness, low coefficient friction, high resistance to oxidation, and high resistance to delamination wear of TiAlN/CrAlN indicating a very promising applicability of this coating for high speed cutting of abrasive wood based materials.

REFERENCES 1.

Stewart H.A., 1992. High-Temperature Halogenation of Tungsten Carbide Cobalt Tool Material when Machining Medium Density Fiberboard. For. Prod. J 42(10):27-31 2. Sheikh-Ahmad J.Y. and Bailey J.A., 1999. High-Temperature Wear of Cemented Tungsten Carbide Tools while Machining Particleboard and Fiberboard. J Wood Science 45 (6):445455 3. Banh T.L., Phan Q.T. and Nguyen D.B., 2004. Wear Mechanisms of PVD Coated HSS Endmills Used to Machine 1045 Hardened Steel. Advances in Technology of Materials and Materials Processing. 6(2) : 244-249. 4. Salje E. and Stuehmeier W., 1988. Milling Particleboard with High Hard Cutting Materials. Proceedings of the 9th International Wood Machining Seminar. pp 211-228 5. Sheikh-Ahmad J.Y. and Stewart J.S., 1995. Performance of Different PVD Coated Tungsten Carbide Tools in the Continuous Machining of Particleboard. Proceedings of the 12th International Wood Machining Seminar. pp 282-291 6. Fuch I. and Raatz Ch., 1997. Study of wear behavior of specially coated (CVD, PACVD) cemented carbide tools while milling of wood-based materials. Proceedings of the 13th International Wood Machining Seminar, pp 709-715 7. Darmawan W., Tanaka C., Usuki H. and Ohtani T., 2001. Performance of coated carbide tools in turning wood-based materials: Effect of cutting speeds and coating materials on the wear characteristics of coated carbide tools in Turning Wood-Chip Cement Board. J Wood Science 47 (5) : 342 - 349. 8. Ding X.Z., Tan A.L.K., Zeng X.T., Wang C., Yue T. and Sun CQ., 2006. Corrosion resistance of CrAlN and TiAlN coatings deposited by lateral rotating cathode arc. Thin Solid Films. 516 : 987-992 9. Chang C.L. , Chen W.C., Tsai P.C., Ho W.Y. and Wang D.Y., 2007. Characteristics and performance of TiSiN/TiAlN multilayers coating synthesized by cathodic arc plasma evaporation. Surface and Coatings Technology, 202 (7) : 987-992 10. Darmawan W., Quesada J., Rossi F., Marchal R., Machi F. and Usuki H., 2009. Performance of laser treated AISI-M2 cutting tool for peeling beech. Eur. J. Wood Prod. 67 : 247-255

96

Tool Material and Tool Wear Poster Presentations


Measurement Of The Cutting Tool Edge Recession With Optical Methods Sandak Jakub1, Palubicki Bartosz2, Kowaluk Grzegorz3 1

Trees and Timber Institute IVALSA/CNR, San Michele All’Adige, ITALY,

[email protected] 2 Poznan University of Life Sciences, Poznan, POLAND 3 Wood Technology Institute, Poznan, POLAND

ABSTRACT The cutting tool wear is a natural phenomenon occurring while machining wood. The tool wear rate, dynamics of the wear progress and mechanisms of recession are very complex and depend on several factors such as; cutting angles, tool material, processed wood properties, machine dynamics, and processing speeds among others. Even if from practical point of view the cutting tool edge geometry is not such important as generated surface roughness, it is crucial for understanding of the wear physics. Consequently it is very important information for tool producers and researchers. Several methods have been applied for estimation of the tool wear. The goal of this work was however to design and verify novel optical instruments (such as laser micrometer and triangulation scanner) in scrutinizing of tool wear and scanning of the cutting tool geometry.

INTRODUCTION Wood cutting is a complex process in which many phenomena interfere together. The cutting tool type and edge geometry, kinematic parameters of machining and material properties decide of machined surface geometry, cutting forces, energy consumption, vibrations, noise level and many others. Sharp cutting edge has a positive impact on widely understood machining quality and therefore it is crucial from industrial point of view to increase the life-span of tool (until it is still sharp enough to ensure quality demands). For this reason new, more wear-resistant coatings [21] and bulk materials are introduced into tool production (e.g. Polycrystalline Diamond, Diamond Dispersed Cemented Carbides). Mechanisms of tool blunting are also considered, Ramasamy and Ratnasingam [18] point that tool wearing might be caused by: gross fracture or chipping (catastrophic) abrasion, erosion, micro fracture, electrochemical corrosion and oxidation. On the other side a worn tool – particularly when catastrophic wear occurred – has a great impact on the surface quality and on dynamic of machine [15]. For all these reasons it is important to recognize the tool wear. It has been of interest of many researchers during years to recognize the tool wear during wood machining. Klamecki [5] states that :“ The change in the cutting tool with use has generally been monitored in two ways, by observing the change in the edge geometry, and by observing changes in the forces acting during cutting”. He also notices that some authors have used other measures of tool dulling e.g. time needed to plane a given length of wood while applying a constant feed force and the resultant deflection of a string before cutting by a worn knife; also a size of sawdust chips has been used as an indicator of tool wear. All the tool wear measurement methods are usually divided into: direct and indirect measures.

97


Direct methods The most obvious way of controlling the tool wear is a direct measurement of tool geometry. It has some disadvantages like time consuming (dismounting, measuring and remounting a tool or a blade) or difficulties in remounting in exactly this same position, but gives an objective and absolute wear data [2]. Three general methods of direct tool tip geometry measurement are utilized: contact, optical, and SEM. Contact methods rely on tip or wedge type stylus moved along the cutting edge (e.g. Miklaszewski et al. [11]) or perpendicularly to achieve tool wedge profile [12]. Optical method has several varieties. In the simplest setup optical microscope is used to observe the tool tip or, if equipped with camera to save the image, analyze afterwards [1, 6, 8]. The variation of this method includes putting Vickers indentation marks [2, 4] or surface scratches [10] on observed face for comparing the distance from the marks to the cutting. In this way the measuring base is constant and the measurement does not depend on the initial edge wear. Rarely – due to high work and time consumption required, a non-destructive silicone cast crosssection method is used to obtain geometry of tool cross-sections [11, 21]. Methods utilizing laser light are also considered as optical methods. Some trials have been undertaken to acquire a measure of tool geometry in-situ on the rotating spindle: Ochuchi et al. [13] have used laser curtain instrument for detecting outer diameter of a router bit. They successfully compared results with a stylus method. Scanning Electron Microscope (SEM) is used to measure the tool wear. In many cases [3, 14, 21] SEM is used in the same manner as an optical microscope to make a high resolution image which are then subjected to geometry measures or simply qualitative analysis. When higher resolutions are used more detailed analysis of wear geometry and mechanisms are analyzed [11, 21]. Regardless of measuring method different parameters are used by authors to quantify tool wear. a)

γF

c)

ra k

e fa

βF

ce

b)

VB

S

β F/2

W VB

αF

e clearan ce fac VBF

Figure 1 Geometrical parameters of the cutting edge wear as considered by McKenzie and Karovich (1975) (a), Porankiewicz et al. (2003) (b) and Sheikh et al. (2003) (c)

98


McKenzie and Karpovich [10] have utilized many parameters (Figure 1a) in order to provide a wide spectrum of tool blunting relations to cutting forces. None of their parameters was linearly related to the cutting forces. Porankiewicz et al. [17] noticed that edge recession on the clearance face (VBF) is in wood cutting the most intensive of all measured (Figure 1b). This is confirmed by Sheikh-Ahmad and Bailey [21] basing on replica cross-section. Sheikh et al. [22] have used nose width – wear land measured in a surface perpendicular to the tool angle bisector (Figure 1 c). Alternatively, a cutting edge rounding (tool tip radius) might be used for quantification of the tool wear [2]. Indirect methods On the other side strong effort is put onto on-line indirect methods of tool wear recognition. From practical point of view it is not the geometry of the tool itself which is important but the effect it makes on the process (quality of generated surface , energy consumption, sound intensity etc). For this reason, as well as because of possibilities of on-line capabilities, indirect measures of tool wear seem to be very attractive alternative for adaptive controls of cutting processes. So far, besides of cutting forces considered by many researchers [1, 5], several other physical effects of cutting have been researched in order to predict the tool wear. These include energy consumption, vibrations [7, 8], acoustic emission [9], and temperature of machined surface [23] among others. It must be mentioned anyway, that for calibration of the indirect measurement systems in most cases detailed information on the cutting tool micro geometry are required as a reference.

OBJECTIVES Even if a lot of research has been dedicated to development of tool wear monitoring techniques there is still, in our opinion, a need to construct a direct cutting edge measurement technique to be utilized for in-line applications. In the principal it should exclude contact with the measured edge. It should allow measurement of the tool without uninstalling it from the spindle (or at least without uninstalling inserts from the tool holder). It should be able to reproduce whole cutting edge length and provide information abut the geometry in three dimensions. Some other minimum requirements to be considered for such scanner could be pointed as bellow: • Minimal radius of the tool tip to be detected: 10µm • Length of the measured edge: up to 30mm • Measurement speed: reasonable for in-line applications (in order of minutes, preferably seconds) • Size of the scanner: compact and rigid • User interface: automatic and intuitive The goal of this project was therefore to develop a novel methodology for estimation of the cutting edge geometry, sensitive enough for industrial applications and with potential for on-line measurements.

METHODS AND RESULTS Several optical techniques have been pre-selected for analysis within the framework of this project: • laser micrometric scanner

99


• • • •

laser displacement sensor laser line triangulation shadow triangulation depth-from-focus

Laser micrometric scanner The successful example of the laser micrometric scanner in estimation of the cutting tool conditions has been reported by Ohuchi et al. [13]. The work presented here was an attempt to extract additional information from their results and to reproduce three dimensional shape of the cutting edge. A dedicated experimental set-up has been developed as presented in Figure 2. The cutting tool n has been installed on the rotary stage q of the scanner. The contour of the tool was scanned by the laser micrometric system o (Mitutoyo LSM-503) while the tool was rotated. The tool (together with rotary stage) was translated horizontally by linear moving stage r in order to scan subsequent sections. The position of the light shadow has been acquired by the micrometer and recorded by computer together with linear and rotational positions.

[

X

\

Z

Y

Figure 2 Experimental platform #1 for scanning cutting edge geometry; tool n, laser micrometer o, laser displacement sensor p, rotary stage q, horizontal moving stage r 3D contour of the whole cutting tool can be scrutinized as a result of scanning with laser micrometric system. An example of such result is presented in Figure 3a. The same 3D map, but represented in the Cartesian space is shown on Figure 3b (the horizontal axis corresponds to rotation angle). The highest points belong to the cutting edges. To visualize an effect on the tool wear two maps scanned from the sharp and extensively used tools are presented in figure 3 (c and d respectively).

100


a)

b)

c)

d)

Figure 3 Results of the diamond tool scanning with laser micrometric system; 3D shadow contour of the tool (a), map of the tool (b), result of sharp tool scanning (c), result of extensively used tool scanning (d) Laser displacement sensor On the same experimental platform #1 as the one used for laser micrometer scanner an additional triangulation laser displacement sensor p was installed. The sensor used was Keyence LKG-32, connected to PC by RS-232. The triangulation is a widely used optical method to calculate a distance between sensor and illuminated spot. In principal the focused spot of laser light is emitted to the surface with a specific (constant) angle. Part of the light is reflected from the surface in to direction of the detector. Analyzing intensities of the pixels illuminating the detector it is possible to calculate distance between surface analyzed and sensor. In case of laser displacement sensor it is possible to estimate only one point, however after rotating and moving the tool (by means of rotary q and linear r stages) it become possible to scrutinize 3D shape of the tool. Even if the resolution of the laser displacement sensor used was superior to the scanning resolution required (10µm), the performance of such system was disappointing. Due to the peculiar geometry of the cutting edge tip (highly polished and curved surface) the laser light was reflected not in to direction of the detector (in diffuse manner), but rather reflected specularly in to direction different than detector. As a result the signal acquired in those areas was an error. In consequence the laser displacement sensor was rejected from further tests. Laser line triangulation Next a series of tests were performed on the surface roughness evaluation experimental platform described in detail in [20]. First attempt was a variation of the laser displacement sensor extended

101


in to line scanning instead of spot. The hardware for the experiment is presented in Figure 4. The source of laser light o (ultra thin laser line projector Stocker Yale Lasiris, 635nm, 1mW) illuminated measured cutting edge installed on the tool n with a structured light (laser line). The image of the laser line shape on the cutting edge was acquired by video camera r (6.6 Mpixels, Pixelink PL-A782) equipped with macro/zero-distortion lenses q (Optoengineering MC3-03X). The focus of the camera and minimization of the laser line with was performed by moving the optical system vertically by means of linear moving stage t. The tool was also moved horizontally with a help of moving stage s in order to generate 3D shape of the cutting edge. A dedicated software controlled all parts of the scanning systems, captured/processed images and generated 3D maps of the cutting edges. An advantage of the laser line triangulation system over laser displacement sensor was radically reduction of the scanning time. Unfortunately, the laser line triangulation method possessed analogous (to laser displacement sensor) limitation – misrepresentation of the cutting edge shape due to extensive laser line width, elevated laser speckle and high specular reflection of the laser light from metallic surface. The signal captured by the camera was very weak and in consequence representation of the cutting edge geometry was poor. It should be mentioned here that it might be possible to improve the set-up (by means of more powerful laser light source or replacing camera by more sensitive one); however it was not considered within the framework of this project.

^ \ Z

[

Y

X ] Figure 4 Experimental platform #2 for scanning cutting edge geometry; tool X, laser line projector Y, shadow projector Z, macro/zero-distortion telecentric lenses [, camera \, horizontal moving stage ], vertical moving stage ^

Shadow triangulation Light section shadow triangulation was also tested as an alternative to laser line triangulation scanner. The hardware used was very similar to the above mentioned (Figure 4); however instead of the monochromatic and highly polarized laser source o, the light section was created by structured light projector p in a form of the half-plane shadow. The source of light (3W LED white light projector LTPR3W/W produced by Optoengineering) was equipped by telecentric lenses. In consequence the scanner hardware was optimized in order to increase the quality of images acquired and reduced, as possible, all optical distortions. Image processing algorithm was

102


developed on the bases of previous work [19], with slight adoption due to peculiarities of the cutting edge. Examples of the 3D scans of the cutting edge (at different stages of wear) are presented on Figure 5. The cutting tool was engaged in processing of particle boards and was measured before and after 500, 1000 and 1500 meters of the cutting distance. Even if the toll was still reasonably sharp at the end of test, clear marks of the wear were observed. As the quality of the results obtained was superior, it was possible to analyze the wear mechanism in detail (such as dynamics of the recession on the rake and clearance faces, etc.). a)

b)

c)

d)

Figure 5 3D reconstruction of the cutting edge geometry after cutting particle board; 0 m (a), 500 m (b), 1000 m (c) e 1500 m (d)

Depth of focus The last technique evaluated was adoption of the depth-from-focus approach (or focus stacking) to reconstruct 3D maps of the cutting edge geometry. The principle of this technique bases on the phenomenon of the focus depth of the lenses. The focal depth of typical lenses is relatively small, and it depends on the quality of optics, numerical aperture and optical magnification. It could be utilized for measurement of the depth, assuming that it is possible to quantify focus numerically. Various algorithms could be applied for focus quantification. The one selected in this project based on maximization of the standard deviation of the pixels intensity. The experimental set-up based on the experimental platform #2 (Figure 4). Images of the cutting edge X at different focal lengths were captured by the camera \. The focal length was controlled by moving vertically camera and lenses [ by means of the vertical moving stage ^. The tool was illuminated by the white light projector Z. A series of blurred images (with different distance of the camera to the edge) were obtained as a result of scanning (Figure 6). Dedicated software post-processed all these images in order to determine the altitude of the maximum focus for all pixels. The “best focus” assessment algorithm applied within this project based on the maximization of the standard deviation of the light intensities in small regions of image; however several alternative algorithms (open source or commercial) are available. It becomes possible to reconstruct the 3D image of the cutting edge basing on the “best focus” information, as presented on Figure 6e. It should be mentioned here that due to not optimized software the image reconstruction of high definition images was rather time consuming (in order of hours). In consequence improvements for the software (and/or hardware) are necessary in order to apply

103


this technique on-line. Beside that, the depth-from-focus method seems to be very effective and relatively accurate. a)

b)

c)

d)

e)

Figure 6 Depth-from-focus measurement of the tool edge; images of the cutting edge corner with four different focal distances (a-d), and resulting 3D map of the edge scrutinized by analyses of these images (gray color corresponds to the depth) (e)

CONCLUSSIONS Basing on the results presented and other experiences acquired during conducted studies the conclusions can be summarized in a form of table 1. Each technique described possesses potential for measurement of the cutting edge geometry; however some improvements to the set-up, hardware and software might be recommended. The most reliable method of all presented here

104


seems to be shadow triangulation. However laser micrometer might be the most suitable for inline application.

performance of prototype

potential for improvement

hardware simplicity

computation/scanning time

3D reconstruction of the cutting edge

on-line potential

Table 1 Summary of the cutting edge geometry measurement techniques researched

laser micrometric scanner

9

9

8

8

8

9

laser displacement sensor

8

8

9

9

9

8

laser line triangulation

8

9

8

9

9

8

shadow triangulation

9

8

8

9

9

9

depth-from-focus

9

9

9

8

9

8

ACKNOWLEDGMENTS Part of this work has been conducted within bi-later Poland - Italy collaboration project and within framework of SWORFISH project co-financed by Provincia Autonoma di Trento.

LITERATURE 1.

2.

3.

4.

5.

Aknouche H., Outahyon A., Nouveau C., Marchal R., Zerizer A., Butaud J.C 2009: Tool wear effect on cutting forces: in routing process of Aleppo Pine wood, Journal of Materials Processing Technology vol. 209: 2918-2922 Bonamini G., Collet R., Del Taglia A., Fibbi F., Goli G., Remorini R., Uzielli L. 1999: Equipement and test method for measuring cutting forces and wear of cutting edges in saw teeth for wood, improved with new HVOF coatings, Proceedings of the 14th International Wood Machining Seminar, September 12-19 1999: 453-461 Fuchs I., Endler I., Peter M 2005: High performance of hard metal tools for woodworking by gas boronizing, Proceedings of the 17th International Wood Machining Seminar, Rosenheim, Germany September 26-28 2005: 542-550 Ishida K., Tsutsumoto T., Banshoya K. 2005: Cutting performance of Edge-sharpened Thick Diamond-film Brazed Milling Tools, Proceedings of the 17th International Wood Machining Seminar, Rosenheim, Germany September 26-28 2005: 550-560 Klamecki B.E. 1979: A Review of Wood Cutting Tool Wear Literature, Holz als Rohund Werkstoff 37:265-276

105


6. 7. 8.

9. 10. 11.

12.

13.

14.

15. 16.

17. 18. 19. 20. 21. 22. 23.

24.

Kowaluk G., Szymanski W., Palubicki B., Beer P. 2009: Examination of tools of different materials edge geometry for MDF milling, Eur. J. Wood Prod. Vol. 67:173-176 Lemaster, R. L.; Lu, L.; Jackson, S. 2000 The use of process monitoring techniques on a CNC wood router. Part 1. Sensor selection. Forest Products Journal vol. 50(7/8):31-38 Lemaster R.L, Lu L., Jackson S. 2000: The use of process monitoring techniques on a CNC wood router. Part 2. Use of a vibration accelerometer to monitor tool wear and workpiece quality, Forest Product Journal vol. 50(9):59-64 Lemaster R., Tee L. 1985: Monitoring tool wear during wood machining with acoustic emission. Wear vol. 101:273-282 McKenzie W.M., Karpovich H. 1975: Wear and blunting of the tool corner in cutting a wood-based material, Wood Sci Technol vol. 9: 59-73 Miklaszewski S., Zurek M., Beer P., Sokolowska A. 2000: Micromechanism of polycrystalline cementem diamond tool wear Turing milling of wod based materials, Diamond and related materials vol. 9: 1125-1128 Novacek E., Novak V. 2006: Possibilities of measuring tool-wear (in Slovak), Medzinárodná vedecká konferencia k 10. Conf Trendy lesnickej, drevarskej a environmmentalej techniky, Zvolen, Slovak Republic, 5. – 7. september 2006: 330 - 336 Ohuchi T., Kameyama J., Murase Y. 2003: Development of automatic measurement system for both wear and cutting edge profile of router bit, Proceedings of the 16th International Wood Machining Seminar, Matsue, Japan, August 24 -27 2003 Okai R., Tanaka C. Iwasaki Y. 2006: Influence of mechanical properties and mineral salts In Wood species on tool wear of high-speed steel and stellite-tipped to ols – consideration of tool wear of the newly developer tip-inserted band Saw, Holz als Rohund Werkstoff vol. 64: 45 - 52 Paris H. and Peigne G. 2007: Influence of the cutting tool geometrical defect on the dynamic behavior of machining, Int J Interact Des Manuf vol. 1: 41-49 Porankiewicz B., Sandak J., Tanaka C. 2003: The HSS tool wear when milling wood, Proceedings of the 16th International Wood Machining Seminar, Matsue, Japan, August 24 -27 2003 Porankiewicz B. Sandak J., Tanaka C. 2005: Factor influencing steel tool wear when milling Wood, Wood Sci Technol vol. 39: 225-234 Ramasamy G., Ratnasingam J. 2010: A review of Cemented Tungsten Carbide tool wear during wood cutting process, Journal of Applied Sciences vol. 10 (22): 2799-2804 Sandak J. and Tanaka C. 2005: Evaluation of surface smoothness using a lightsectioning shadow scanner, Journal of Wood Science 51(3):270–273 Sandak J. 2007: Optical triangulation in wood surface roughness measurement, Proceedings of the 18th International Wood Machining Seminar, Vancouver, Canada Sheikh-Ahmad J.Y., Bailey J.A. 1999:The wear characterisctics of some cemented tungsten carbides in machining particleboard, Wear vol. 225-229: 256 - 266 Sheikh-Ahmad J.Y., Stewart J.S., Feld H. 2003: Failure characteristic of diamond coated carbides in machining wood-based composites, Wear vol. 255: 1433-1437 Sokolowski W., Gogolewski P. 1999: Temperature of machined surface as a value for tool condition monitoring during woodproducts milling, Proceedings of the 14th International Wood Machining Seminar, Paris-Epinal-Cluny, France Tanaka C., Nakao T., Nishino Y., Hamaguchi T., Takahashi A. 1992: Detection of wear degree of cutting tool by acoustic emission signal. Mokuzai Gakkaishi vol. 38(9): 841846

106


Wear of Teeth of Circular Saw Blades Mats Ekevad, Birger Marklund, Luís Cristóvão Luleå University of Technology, Skellefteå, Sweden [email protected]

ABSTRACT Measured wear data is presented for 3 different carbide grades. The data were collected during rip sawing wood with a double arbour saw. The purpose of the test was to determine the suitability of different grades for sawing frozen timber. A set of circular saw blades of diameter 350 mm was equipped with teeth comprised of 3 different cemented carbide grades, denoted A, B and C. These 3 grades were chosen out of 6 grades in an earlier preliminary laboratory test. Grade A was a relatively soft and tough standard grade (K1C=17, HV30=1150) normally used in sawmills in winter conditions for frozen wood, grade B was harder and more brittle (K1C=11, HV30=1600) and grade C was even harder and more brittle (K1C=9.5, HV30=1950). The double arbour saw (with vertical arbours) was equipped with 6 saw blades (3 on each arbour) for cutting 2 centre boards and 2 side boards. The 6 saw blades with different teeth were mounted in a mixed manner on the arbours, and after sawing a number of logs the wear of teeth was measured. After some time the blades were removed, inspected, ground and used again for sawing. The thickness of boards was also measured and the standard deviation was calculated. The results show cutting edge radii as a function of the number of logs sawn and also the standard deviation of the thickness of the sawn boards. Grade A had the highest wear and grades B and C the lowest wear. There was no significant edge damage during the tests. Grade C did not suffer problems of chipping from cutting edges and was found to be suitable for sawing frozen timber. The thickness standard deviations were constant at about 0.2 mm, and not a function of the number of logs sawn.

INTRODUCTION Circular saw blades with cemented tungsten carbide teeth are normally used in Swedish saw mills for the rip-sawing of logs. Different kinds (grades) of cemented carbides can be used, from softer and tougher ones to harder but more brittle ones. Winter conditions in Scandinavia, in combination with green logs which contain much icy sapwood, are particularly challenging when it comes to the choice of which cemented carbide grade to use. Hard grades wear less but are more brittle, and softer grades wear faster but are tougher. Greater brittleness results in greater damage to the teeth edges, such as corners or small parts (chips) of the edges breaking off. Wear of cutting edges in general has been studied extensively and is caused by a combination of chemical, physical and mechanical processes [1- 9]. Mechanical wear is due to high stresses, friction and partly due to hard particles imbedded in the wood. Chemical wear is due to chemical content in the wood which reacts with the edge material. High temperatures affect

107


wear, since mechanical and chemical properties and reactions are greatly affected by temperature. Wear can be measured by measuring the edge radius, see Fig. 1. Also by inspecting the edge in order to identify broken off parts of the edge.

Center boards

Fig. 1. Definition of edge radius

Fig. 2. Saw blade positions in the double arbour saw

In this study, three different cemented carbide grades are compared. The three grades were tested in circular saw blades used in normal production in a Swedish saw-mill under winter conditions. The purpose was to decide whether it was possible to use harder carbide grades during winter sawing than those normally used. Edge wear, edge damage and the accuracy of board dimensions were registered as a function of the number of logs that were sawn.

MATERIAL AND METHODS A double arbour saw (with vertical arbours) was used for the tests. It was run under normal saw mill production but equipped with test saw blades in different configurations. It had 6 saw blades (3 on each arbour) for cutting 2 centre boards and 2 side boards, see Fig. 2. The geometry of the saw blades that were used are shown in Fig. 3. The diameter was 350 mm, the thickness was 2.6 mm, the kerf width was 3.6 mm, the number of teeth was 33, the rake angle was 30°, the clearance angle was 10° and the rotation speed was 3300 rpm. Four teeth on each tested saw blade was marked and checked after sawing a number of logs. The mean edge radii of the 4 teeth were measured and the 4 teeth were photographed. The blades were then put back into production. After an additional period of time the blades were removed, inspected, ground and used again for sawing. The thickness of the sawn boards was measured at the top end, root end and in the middle of 10 boards for each occasion. The standard deviation of the thickness was calculated.

108


Standard deviation (mm)

0.3 0.25 0.2 0.15 0.1 0.05 0 0

2000

4000

6000

8000

Number of logs

Fig. 3. Saw blade geometry

Fig. 4. Each point is the standard deviation (mm) for thickness at 3 positions for 10 centre boards

The normal grade used in Sweden for winter-sawing is relatively soft and tough and was here denoted grade A, with a fracture toughness K1C=17 MNm-3/2 and hardness 1150 HV30. Grade B was harder and more brittle, K1C=11 MNm-3/2, hardness 1600 HV30. Grade C was even harder and more brittle K1C=9.5 MNm-3/2, hardness 1950 HV30. Four tests were conducted from March 2009 until August 2010 with the saw blades of different teeth material in different positions, see Table 1. Both pine and spruce logs were sawn, in both frozen and unfrozen conditions. Blade numbers 1 to 3 were equipped with grade A teeth, blade numbers 4 to 6 had grade B teeth and blades 7 to 9 had grade C teeth.

RESULTS The standard deviations for board thicknesses are shown in Table 2 and Fig. 4. The edge radii are shown in Table 3 and Fig. 5. Edge damage was not found after inspection after normal sawing in test 1 and test 5, except after test 1 where minor edge damage (small chips broken off) were found for grade B. Accidents that stopped test 2 and test 4 (jammed log in sawing machine) resulted in severe blade failure but were not of interest in this study.

109


Table 1. Description of tests Test 1 2009-03-09 Start date (spruce), then 200905-20 (pine) Spruce+pine Log material Frozen spruceLog condition unfrozen pine Feeding speed 70 (spruce) 55 (pine) (m/min) Centre board 47x150 (spruce) 63x155 (pine) dimension B No 4 Pos. 1 A No 1 Pos. 2 B No 5 Pos. 3 A No 2 Pos. 4 B No 6 Pos. 5 A No 3 Pos. 6 Number of logs 5 spruce sawn at first stop Number of logs 2908 spruce sawn at second stop Number of logs sawn at third stop Number of logs sawn at fourth stop Action after test

5062 spruce

Test 2 2009-10-05

Test 4 2010-03-18

Test 5 2010-08-20

Pine Unfrozen

Spruce Frozen

Spruce Unfrozen

76

66

68

50x150

63x150

50x150

B No 4 A No 1 B No 5 A No 2 B No 6 A No 3 5

C No 7 A No 1 B No 5 C No 8 B No 4 A No 2 5

C No7 A No 3 B No 4 C No 8 B No 6 A No 1 5

5303

3120

2870

5303+150

5062 spruce+1607 pine (2009-05-20) Regrinding. Accident+ Minor edge regrinding damages in grade B

Accident+ regrinding of all except the grade C blades that were undamaged

Table 2. Number of logs sawn/ standard deviation of thickness of boards (mm) Test 1 Test 2 Test 4 Test 5

First stop 5/0.21 5/0.28 5/0.23 5/0.18

Second stop 2908/0.16 5303/0.26 3120/0.18 2870/0.24

Third stop 5062/0.18

110

Fourth stop 5062+1607/0.21


Table 3. Edge radii (μm), grade of teeth (A, B or C)

Test 1

stop 1 stop 2 stop 3 5062 logs stop 4 stop 1 stop 2 5303 logs stop 3 stop 1 5 logs stop 2 3120 logs stop 1 stop 2 2870 logs

Test 2

Test 4

Test 5

Pos. 1

Pos.2

Pos.3

Pos.4

Pos.5

Pos.6

40 B

52 A

38 B

54 A

69 B

55 A

60 A

35 B

20 C

16 A

30 B

20 C

30 B

16 A

26 C

27 A

33 B

26 C

33 B

27 A

43 C (not 48 A reground)

43 B

43 C (not 43 B reground)

48 A

70 Unfrozen pine test 2

Grade A Grade B

60

Unfrozen spruce test 5

Grade C

Edge radius ( m)

50

Frozen spruce test 1

Unfrozen spruce test 5

40

30 Frozen spruce test 4

20 Before test 4

10

0 0

1000

2000

3000

4000

5000

Number of logs

Fig. 5. Edge radii of teeth (μm) from all tests

111

6000

7000


ANALYSIS AND DISCUSSION Fig. 4 and Table 2 show that the standard deviation of the thicknesses of the 10 measured boards was about 0.2 mm, and that this value was constant and not affected by the number of logs sawn. Fig. 5 and Table 3 show that edge radii for all grades in general increased with the number of logs sawn. Also that grade B and C resulted in the least amount of wear. Grade A presented the greatest wear as was expected before the tests. Since the data points were taken from different tests conducted on different occasions, for both pine and spruce logs and for both frozen and unfrozen wood, the edge radii did not consistently increase with the number of logs sawn. The edge radii at the start after grinding were not measured except for test 4, and were possibly different in the other tests. There was no significant edge damage to the cutting edges (chips broken off) for any of the grades.

ACKNOWLEDGEMENT The authors express their gratitude to the European Regional Development Fund, Objective 2, Northern Sweden via Tillväxtverket (the Swedish Agency for Economic and Regional Growth) and Vinnova (the Swedish Agency for Innovation Systems) for financial support.

REFERENCES 1.

Klamecki, B. (1979) A review of wood cutting tool wear literature. Holz Roh Werkst 37:265–276.

2.

Kirbach, E. & Chow, E. (1976) Chemical wear of tungsten carbide cutting tools by western red cedar. Forest Prod. J 26(3):44–48.

3.

McKenzie, W. & Cowling, R. (1971) A factorial experiment in transverse-plane (90/90) cutting of wood. Part I. Cutting force and edge wear. Wood Sci. 3: 204–213.

4.

McKenzie, W. & Karpovich, H. (1968) The frictional behavior of wood. Wood Sci. Technol. 2: 138–152.

5.

Okai, R. & Tanaka, C. & Iwasaki, Y. (2005) Influence of mechanical properties and mineral salts in wood species on tool wear of high-speed steels and satellite-tipped tools – Consideration of tool wear of the newly developed tip-inserted band saw. Holz Roh Werkst 64:45–52.

6.

Porankiewicz, B. & Grönlund, A. (1991) Tool wear-influencing factors. In Proceedings of the 10th International Wood Machining Seminar. 21-23 October, University of California, Richmond. pp. 220–229.

7.

Porankiewicz, B. & Iskra, P. & Sandak, J. & Tanaka, C. (2006) High-speed steel tool wear during wood cutting in the presence of high-temperature corrosion and mineral contamination. Wood Sci. Technol. 40:673–682.

112


8.

Porankiewicz, B. & Sandak, J. & Tanaka, C. (2005) Factors influencing steel tool wear when milling wood. Wood Sci. Technol. 39:225–234.

9.

Sheikh-Ahmad, J. & McKenzie, W. (1997) Measurement of tool wear and dulling in the Machining of particleboard. In Proceedings of the 13th International Wood Machining Seminar. 17-20 June, Vancouver, Canada. pp. 659–670.

113


Agricultural Residues in Panel Production – Impact of Ash and Silica Content Müller, Christian1; Deetz, Richard2; Schwarz, Ulrich1; Thole, Volker2 1

Eberswalde University for Sustainable Development - University of Applied Sciences, Faculty of Wood Science and Technology, Alfred-Möller-Str. 1, 16225 Eberswalde, Germany 2

Fraunhofer-Institute for Wood Research, Wilhelm-Klauditz-Institut (WKI), Bienroder Weg 54E, 38108 Braunschweig, Germany

ABSTRACT Considering ongoing global deforestation and the increased demand for raw materials in all wood-processing sectors it is necessary to seek and employ alternative resources. Particularly for the wood-based panel industry even partial substitution of conventional wood material becomes crucial. Agricultural residues, for example cereal straws or suitable annual fibres, are potential sources of alternative lignocellulose-based raw materials. Research on substitute materials for wood panel production mainly focused on physical and mechanical properties of boards produced from these materials while much remains to be learned about machining. Therefore studies were carried out to analyse the interrelationship between ash and silica content and their impact on tool wear. Ash and silica contents of selected agricultural residues at different phenological development stages and from different parts of the examined plants were determined following ISO 3340. Furthermore, tool wear tests were carried out using particleboards made from wheat straw, canola straw, and spruce as well as commercial raw particleboards. Analyses of various plant materials displayed no signs of a functional interrelationship between ash and silica content. Preliminary results of the tool wear tests show an increased edge recession for the agro-based particleboards compared to a reference sample made from spruce while compared to commercial raw particleboard the wear was lower or similar. The evaluation of the plant material confirmed that it is not possible to predict silica content of composite panels by measuring just their ash content. Analyses of wear behaviour indicate that the impact of abrasive substances on edge recession is strongly influenced by the raw density of the panels. Furthermore, it seems that tool wear is less affected by biologically deposited silica in the cell walls of agro-based lignocelluloses than by contaminations with coarse silica from soil.

INTRODUCTION Besides wood as most common raw material for particle and fibre boards other lignocellulosebased materials are also suitable for panel production. The utilization of agricultural residues as a raw material for pulp and panel production is not a new approach. With changing intensity it has been a matter of economic and scientific interest since the first decades of the 20th century [1-3]. During the last decades the demand for raw materials in all wood-processing sectors and also for wood-based panels has increased, causing a worldwide shortage and decline in forest resources followed by increasing costs for wood. For the future the same trend is forecasted [4, 5]. To face the resulting impacts of this development on the economical, ecological and socio-economic situation it is necessary to seek and employ alternative resources to substitute wood.

114


The development of composite panels based on agricultural residues is of increasing importance. These different materials represent renewable, environmentally friendly resources. They have the ability to satisfy the increasing demand for lignocellulose-based raw materials, even if it is borne in mind that just a smaller proportion of the total volume is available for material utilization. The utilization reduces the pressure on the forest resources as well as the environmental impacts caused by disposal of residues (field burning, water and land-consumptive decomposition). Furthermore, it can provide farmers an additional income from cultivation of crops [1]. In recent years the development of marketable particle and fibre boards made from different raw materials has been in the focus of many research projects. A compilation of research papers and reports is provided by Müller et al. [6]. The most frequently considered non-wood materials are cereal straws, such as wheat or rice straw, bagasse, kenaf, bamboo, and palm residues. Furthermore maize, hemp, flax, cotton as well as different kinds of reed and gramineous fibres were examined. In addition also materials like stalks from sunflower, castor-oil plant, tomato, eggplant, different nutshells, and prunings were analysed in regard to their suitability. A comprehensive survey of research studies on the use of non-wood plant fibres (period 1913-1993) provided Youngquist et al. [7]. Besides their suitability as wood substitute specific annual fibre plants provide a high innovation potential for special applications. Based on low bulk density, it is possible to produce panels with lower densities than conventional wood-based panels [8, 9] or with improved insulation properties [10, 11]. It is also possible to improve the mechanical properties due to a greater mean compression ratio of the low density material [12] or by using specific materials with high fibre strength as reinforcement [13, 14]. The higher silica contents compared to wood reduces the inflammability of these materials [15, 16]. Compared to wood different aspects and possible additional costs have to be taken into account while using agricultural residues. Especially in temperate zones the harvesting season is short. In subtropical and tropical regions a better supply is possible due to longer growing seasons respectively additional harvest cycles. Most agro-based residues have low bulk densities. Therefore compaction (bailing) and high storage capacities are required to ensure a continuous production. During storage adequate protection must be provided to prevent biological degradation and contamination [17]. To avoid excessive costs for supply, transport and storage processing in small scale mills close to rural areas is advantageous [1, 2]. Agricultural residues may be difficult to bond with conventional adhesives like ureaformaldehyde (UF) resins. Agro-based lignocelluloses contain high amounts of extractives [18, 19] as well as wax-like substances and silica [17]. They influence the wettability and the gel time of UF resin due to their high buffering capacity and result in poor bond strength [20, 21]. Alternative adhesives such as isocyanates can be used for high quality panels, but have attendant disadvantages (higher cost, release agent etc.). For manufacturing panels fulfilling the standard requirements using UF resins it is necessary to modify the properties of the raw material. Modification due to bleaching, application of coupling agents or enzymes, steam-explosiontreatment and other chemi-thermo-mechanical pre-treatments can improve the board properties [2, 22-26]. Up to now research on substitute materials for panel production mainly focused on physical and mechanical properties of boards. In regard to tool wear behaviour there are just few data available, although this is of particular importance. Commonly, it is derived from the ash and silica content of the materials used. Annual and perennial lignocelluloses show noticeably higher ash and silica contents than wood [3, 27]. Therefore increased tool wear is predicted.

115


MATERIALS & METHODS Subject of the study was the interrelationship between ash and silica content at different phenological development stages and from different parts of the examined plants. Material has been taken from amaranth (Amaranthus sp.), buckwheat (Fagopyrum esculentum), barley (Hordeum vulgare), canola (Brassica napus), hemp (Cannabis sp.), Jerusalem artichoke (Helianthus tuberosus), maize (Zea mays), millet (Panicum miliaceum), rye (Secale cereale), sudan grass (Sorghum xdrummondii), sunflower (Helianthus annuus), and wheat (Triticum aestivum). The plant material was washed, dried and separated into stalks, leafs and infructescences (without seeds or grains). The ash and silica contents were determined according to ISO 3340 [28] using a microwave muffle furnace. Deviating from ISO 3340 the weight (oven dry) of the samples was 10...15 g, they were digested in hypochloric acid (20 ml, 18.5 %) and boiled up. Afterwards, the samples were washed and decanted twice with distilled water (20 ml) and dried in a muffle furnace to avoid filter loss (particles smaller than 40 µm). Furthermore, the tool wear was examined after machining particleboards (PB) made from selected annual fibre plants with high (wheat straw) and low (canola straw) silica contents. As reference samples PB made from spruce (Picea abies) and commercial three-layer raw particleboard were used. A series of experimental single-layer boards (resin content = 5 % polymeric diphenylmethane diisocyanate (pMDI), raw density = 0.65 g/cm3, board thickness = 16 mm) was made under laboratory conditions. The routing of board specimens was performed with a CNC router under the following conditions: standard carbide blades (ISO-Code: K05); wedge angle = 55°; number of cutting edges = 1; cutting speed = 67 m/s; feed rate per tooth = 0.65 mm; feed rate = 10.4 m/min; cutting depth = 3 mm. The tool wear (edge recession) was estimated at regular intervals using a tactile measuring device (Mahr Perthen FRW-750). Edge recession caused by the middle layer (ML) is given as area of material loss to ensure the comparability of all different panels examined.

RESULTS & DISCUSSION The analyses of the plant material did not show any direct functional interrelationship between ash and silica content. Plants with high ash contents do not necessarily have high silica contents and vice versa (Fig. 1, 2). The species-specific ash and silica contents vary with the phenological development of the plants and with the plant components (stalks, leafs, infructescences). Distinctive differences were found; leafs contain much more ash and silica than stalks (Fig. 1, 2). Furthermore, the ash contents in most stalks tend to decline with progressing phenological development stage while silica contents increase. Due to the high silica contents in leafs and parenchyma it is advisable to remove those parts before processing. Plants with low silica content in the stalks were amaranth, buckwheat, canola, hemp and Jerusalem artichoke while wheat straw contained the highest portion. Thereby, it has to be kept in mind that plant silica content can vary because it is influenced by location factors (soil, climate) as well as age, season et cetera. Preliminary results of the tool wear tests after a cutting path of 10,000 meters are displayed in Table 1 and Figure 3. As anticipated, the particleboards made from wheat and canola straw showed increased edge recession compared to spruce PB. An unexpected result in this context was that the canola PB with low silica content caused higher tool wear than the wheat PB with a silica content which is eight times higher. A possible explanation for this phenomenon is a variance in the density profile of the canola PB. This hypothesis is supported by the different wear patterns of the blades. While a slightly higher density in the surface layer of wheat PB and spruce PB can be assumed, the wear pattern of the canola PB indicate a lower density of the

116


surface layer. In case of an average density of 0.65 g/cm3, this means that the middle layer has a higher density than the other boards. In order to verify this hypothesis more vertical density profile tests will be conducted with samples taken from different locations of the board. The commercial raw PB showed a high abrasiveness. The edge recession of the middle layer was three times higher than the wear caused by spruce1. The abrasiveness and increased silica content were probably induced by low-quality raw material. This assumption is based on the presence of plastic waste, bark and also particles of hardwood in the particleboards. Therefore, it can be assumed that waste wood was used for panel production. For the experimental boards a contamination with sand can be excluded because of the pre-treatment of the raw material. No strong correlation was found between silica content and edge recession. Although the agrobased PBs contain much higher silica contents than the wood-based PBs, their edge recession is lower or similar compared to the commercial raw PB. The increased wear behaviour of the commercial raw PB can be explained by the occurrence of coarse sand particles. However, the complexity of influences has to be kept in mind. At present not all differences can be clearly attributed to single factors.

CONCLUSION & OUTLOOK The analyses of the plant material show that it is not possible to predict silica content of composite panels by measuring their ash content only. Using this shortened procedure, like occasionally observed in practice, leads to inadequate assessments. Further analyses of the ash concerning the content of acid-insoluble fractions are necessary. These results and conclusions are confirmed by comparable studies related to wood-based panels [29] and solid wood [30]. The results from tool wear tests indicate that the impact of abrasive substances on edge recession is overlaid by the raw density of the panels. Furthermore, it is assumed that the molecular dispersion of silica in the cell walls of agricultural residues has a different influence on tool wear than contaminations with coarse silica caused by harvest and storage. This assumption, earlier discussed by Sauter [31], would mean that the silica content (without knowing the grain size) as main indicator for wear behaviour is overrated. Work is in progress to quantify and verify the data on tool wear for different panel compositions. Furthermore, it is planned to differentiate the impact of influential factors (silica content, silica size, density, particle size etc.).

ACKNOWLEDGEMENT Financial support by the German Federal Ministry of Education and Research (BMBF) via the German Federation of Industrial Research Associations (AiF) under grant 1723X08 is gratefully acknowledged. Furthermore the authors express their gratitude to AKE Knebel GmbH & Co. KG and TIGRA GmbH for their contribution to this project and the delivery of test materials.

1

In this case the corresponding effect of the surface layer was not considered because of its incomparability (higher density, smaller particle size etc.).

117


Ash and silica content of stalks

30 25 20 15 11.67

10

10.77 10.40 8.55 6.86

5.76 3.69

5 0.37

7.04

0.17

5.60 3.83

0.18

5.11

5.48

3.39

0.25

0.40

1.28

1.67

Ash content Silica content

5.19

2.21

2.25 0.22

A

m ar an t Ba h Bu rl ck ey w he C at Je an ru ol sa a le m He ar mp tic ho k M e ai ze M ill et Su d a Ry n e Su gra nf ss lo w e W r he at

0

Fig. 1: Ash and silica content (stalks) of examined agro-based lignocelluloses

Ash and silica content of leafs

30

27.90

25 21.21

21.37

20 16.12

15

Ash content Silica content

12.95

10 6.24

10.33 8.25 6.06

3.96

5

9.43 5.88 3.35

1.18

A

m ar an t Ba h Bu rl ck ey w he C at Je an ru ol sa a le m He ar mp tic ho k M e ai ze M ill et Su d a Ry n e Su gra nf ss lo w e W r he at

0

Fig. 2: Ash and silica content (leafs) of examined agro-based lignocelluloses

118


0

2

4

6

8

10

12

14

16

0,00 -0,01

Edge recession

-0,02 -0,03 -0,04 -0,05 -0,06 -0,07 Wheat straw PB

Canola straw PB

-0,08 -0,09

Surface layer

Middle layer

Surface layer

-0,10

Width of contact 0

2

4

6

8

10

12

14

0,00 -0,01

Edge recession

-0,02 -0,03 -0,04 -0,05 -0,06 -0,07 Spruce PB

Commercial raw PB

-0,08 -0,09

Surface layer

Middle layer

Surface layer

-0,10

Width of contact

Fig. 3: Wear patterns caused by different particleboards (PB)

119

16


Tab. 1: Edge recession, ash and silica content of examined particleboards (PB) Wheat straw PB

Canola straw PB

0.154

0.302

0.091

0.272

Ash content [%]

3.45

3.47

0.27

0.72

Silica content [%]

1.75

0.22

0.03

0.11

Edge recession [mm2] of ML

Spruce PB Commercial raw PB

REFERENCES 1. Çöpür, Y.; Güler, C.; Akgül, M.; Taşçıoğlu, C. (2007) Some chemical properties of hazelnut husk and its suitability for particleboard production. Build Environ 42: 2568-2572. 2. Thole, V. (2001) Faserplatten aus Palmenresten. Holz + Kunststoff 4/2001: 90-92. 3. Youngquist, J.A.; English, B.E.; Spelter, H.; Chow, P. (1993) Agricultural fibers in composition panels. In: Proc. 27th Particleboard/Composite Materials Symposium, March 3031, April 1, Pullman, USA. 4. Neufeld, B. (2010) Particleboard and medium density fibreboard in the Pacific Rim and Europe 2009 – 2013. BIS Shrapnel Business Research and Forecasting, Sydney. 5. UNECE/FAO (2010): Forest products annual market review 2009-2010. Geneva Timber and Forest Study Paper 25, United Nations, New York and Geneva. 6. Müller, C.; Schwarz, U.; Thole, V. (2011) Zur Nutzung von Agrar-Reststoffen in der Holzwerkstoffindustrie. Holz Roh Werkst (submitted). 7. Youngquist, J.A.; English, B.E.; Scharmer, R.C.; Chow, P.; Shook, S.R. (1994) Literature review on use of nonwood plant fibers for building materials and panels. Gen. Tech. Rep. FPL-GTR-80, USDA Forest Serv., Forest Prod. Lab., Madison, WI, p 146. 8. Dix, B.; Meinlschmidt, P.; van de Flierdt, A.; Thole, V. (2009a) Leichte Spanplatten für den Möbelbau aus Rückständen der landwirtschaftlichen Produktion - Teil 1. Holztechnologie 50 (2): 5-10. 9. Mo, X.; Hu, J.; Sun, X.S.; Ratto, J.A. (2001) Compression and tensile strength of low-density straw-protein particleboard. Ind Crop Prod 14: 1-9. 10. Khedari, J.; Nankongnab, N.; Hirunlab, J.; Teekasap, S. (2004) New lowcost insulation particleboards from mixture of durian peel and coconut coir. Build Environ 39 (1): 59-65. 11. Richter, C. (1993) Neues Verfahren zur Herstellung von Dämmstoffen niedriger Dichte aus Holz und Einjahrespflanzen. Holz Roh Werkst 51: 235-239. 12. Moslemi, A.A. (1974) Particleboard, vol 1: materials. Southern Illinois Uni. Press, Carbondale & Edwardsville. 13. Munawar, S.S.; Umemura, K.; Kawai, S. (2007) Characterization of the morphological, physical, and mechanical properties of seven nonwood plant fiber bundles. J Wood Sci 53: 108-113. 14. Tröger, F.; Barbu, M.C.; Seemann, C. (1995) Verstärkung von Miscanthus- und Holzspanplatten mit Flachsfasermatten. Holz Roh Werkst 53: 268. 15. Thole, V. (2005) Einjahrespflanzen als Rohstoff für MDF - aktueller Stand. MDF Magazin 11: 38-43. 16. Youngquist, J.A.; Krzysik, A.M.; English, B.W.; Spelter, H.N.; Chow, P. (1996) Agricultural fibers for use in building components. In: The use of recycled wood and paper in building applications. Proc. No. 7286, Madison, WI: Forest Products Society.

120


17. Mantanis, G.; Nakos, P.; Berns, J.; Rigal, L. (2000) Turning agricultural straw residues into value-added composite products - A new environmentally friendly technology. In: Proc. 5th Int. Conference on Environmental Pollution, August 28-31, Thessaloniki, Greece. 18. Guntekin, E.; Karakus, B. (2008) Feasibility of using eggplant (Solanum melongena) stalks in the production of experimental particleboard. Ind Crop Prod 27 (3): 354-358. 19. Abdul Khalil, H.P.S.; Siti Alwani, M.; Mohd Omar, A.K. (2006) Chemical composition, anatomy, lignin distribution, and cell wall structure of malaysian plant waste fibers. BioResources 1 (2): 220-232. 20. Hse, C.-Y.; Kuo, M.-L. (1988) Influence of extractives on wood gluing and finishing—a review. Forest Prod J 38 (1): 52–56. 21. Han, G.; Umemura, K.; Zhang, M.; Honda, T.; Kawai, S. (2001) Development of highperformance UF-bonded reed and wheat straw medium-density fiberboard. J Wood Sci 47: 350-355. 22. Han, G.; Umemura, K.; Kawai, S.; Kajita, H. (1999) Improvement mechanism of bondability in UF-bonded reed and wheat straw boards by silane coupling agent and extraction treatments. J Wood Sci 45: 299-305. 23. Zhang, Y.; Lu, X.; Pizzi, A.; Delmotte, L. (2003) Verbesserung der Bindung bei Weizenstrohplatten durch Vorbehandlung mit Enzymen. Holz Roh Werkst 61(1): 49-54. 24. Han, G.; Deng, J.; Zhang, S.; Bicho, P.; Wu, Q. (2010) Effect of steam explosion treatment on characteristics of wheat straw. Ind Crop Prod 31: 28-33. 25. Mantanis, G.; Berns, J. (2001) Strawboard bonded with urea-formaldehyde resins. In: Proc. 35th Int. Particleboard/Composite Material Symposium, April 2-5, Pullman, USA. 26. Markessini, E.; Roffael, E.; Rigal, L. (1997) Panels from annual plant fibers bonded with ureaformaldehyde resins. In: Proc. 31th Int. Particleboard/Composite Materials Symposium, April 8-10, Pullman, USA. 27. Widyorini, R. (2005) Self-bonding characterization of non-wood lignocellulosic materials. Dissertation, Kyoto University. 28. ISO 3340-1976 Fibre building boards – Determination of sand content. 29. Schriever, E.; Boehme, C. (1984) Bestimmung von mineralischen Bestandteilen in Spanplatten. Holz Roh Werkst 42: 51-54. 30. Torelli, N.; Čufar, K. (1995) Mexican tropical heartwoods. Comparative study of ash and silica content. Holz Roh Werkst 53: 61-62. 31. Sauter, S.L. (1996) Developing composites from wheat straw. In: Proc. 30th Int. Particleboard/Composite Materials Symposium, Pullman, USA.

121


122

5. Process Monitoring Oral Presentations


Cant Shape Measurement System and Knife Wear Tests Laganière, Benoît FPInnovations, Québec, QC, CANADA ABSTRACT The cant shape analyzer is an equipment to measure the geometry and surface quality of cants during lumber production at the first or the second log breakdown. It measures cant width, sawing variation, cant taper, log rotation, log centering, cant surface quality and cant face parallelism. The equipment has been installed in four sawmills in Eastern Canada over a period of 18 months. The sawing process is variable and many unexpected problems were detected because cant width varied (decrease or increase) more than 0.050 inch (1.27 mm) in less than 2-3 hours of operation and then remained stable for many hours. Money losses due to different problems are estimated to more than 1 million $ CDN and can be recovered by the installation of this system. Knife wear was evaluated with the system: the overall conclusion is that saw filers change the knives too often. Treatment with a plasma reactor improved knife wear resistance from 35 to 43%.

INTRODUCTION Quality control in sawmills is done manually at different moments by the quality controller, which measures the thickness and width of dry boards after sawing and planing. Quality control is rarely done on cants and the low quantity of pieces measured does not provide an overall representation of the situation of problems that can occur during production. FPInnovations acquired in 2009 a mobile cant analyzer from CRIQ (Centre de recherche Industrielle du Québec) to identify problems and assess the situation of the primary processing industry. Four mills have been studied over a period of 18 months in Eastern Canada. When primary log breakdown is poorly done, wood is lost and unavailable for the successive stages of production. Continuous measuring of cants is the best solution to avoid problems at this important stage of production.

EXPERIMENT Cant shape measurement The cant shape analyzer, developed by CRIQ, is a vision system that can measure surface quality and log positioning precision during primary log breakdown.The analysis software allows real time monitoring of the breakdown process, lumber size (cant size), manufacturing defects, positioning accuracy (log rotation, log centering and cant face parallelism) and cant surface quality (roughness and fiber tear-out).

123


A database provides results for different periods and presents results in through a graphical interface. The performance level of this equipment is ensured by precise measurements, the high level of data aquisition and treatment in real time. The high number of pieces sampled and observation duration provide a solid basis to compare performance over time. To accomplish these functions, the system registers measurements perpendicularly to the axis of the log (log diameter, height and width). The profile of the faces and their position in space are obtained by laser triangulation. A laser line is projected transversely and illuminates the side of the cant face to be measured. A camera captures the laser reflection.

Figure 1 View of the cant shape analyzer baseframe and log scanning

Figure 2 Acquisition of the cant profile by laser cameras

124


The cant shape analyzer is equipped with two laser cameras to acquire cant surface quality data on each side of the cant. The laser, on one side of the cant, projects a thin laser line that is read by the camera, which is called a plan. Each plan is read every 0.5 inch throughout the log. Plan profiles are analyzed and data such as surface tear-out and tear-out depth can be recovered in an Excel file. The following figure shows the mean width variation and standard deviation over one week of production where each point represents an average of 25 logs.

Figure 3 Screen capture – Mean width variation and standard deviation Effect of rotation Average log rotation errors are usually low for all log diameters The mean error was sometimes a few degree negative for a product and a few degree positive for another, which explains the very low mean error (positive or negative). To limit the number of simulations, the average rotation errors were considered null. Figure 4 illustrates monetary losses as % according to different rotation standard deviations encountered during the study where the majority is between 20 ° and 40° that have led to monetary losses below 1%. In conclusion, despite a high standard deviation of 60 °, the monetary losses remained low at 1.5%. We can therefore conclude that rotation errors in a process lead to monetary losses below 2%.

125


Monetary losses (%)

2.0

1.5

1.0

0.5

0.0 0

20

40

60

Rotation standard deviation (deg)

Figure 4 Effect of rotation standard deviation variation on monetary losses Effect of off-centering Figure 5 shows the losses associated with the average centering error and the effect of mean centering errors standard deviation. As for rotation errors, the average centering error varies mainly of a few thousandths of an inch in positive or negative. Manual tests on some samples have produced average centering errors of over 0.300 inch. For a given standard deviation, we see an increase in mean losses of less than 1% by 0.100 inch off centering. Losses increase drastically when the average centering error exceeds 0.300 inche. The graph also shows the effect of centering error standard deviation, which increases by almost 2% for each increase of 0.100 inch (each curve represents a different standard deviation). The results also rarely showed standard deviations lower than 0.100 inch while the average is between 0.100 and 0.200 inch. Standard deviations exceeded 0.300 inch products and resulted in monetary losses exceeding 4%. Losses caused by centering errors can be very important for sawmills if they are not controlled. Similarly, the standard deviation should be decreased. Losses of 2 to 4% of total revenues can be easily caused by poor control of log centering.

126


12

Monetary losses (%)

10

8

6 STD (inch) 0,1 0,2 0,3 0,4

4

2

0 0.050

0.100

0.200

0.300

0.400

Mean centering error (in.)

Figure 5 Effect of mean centring error and standard deviation on monetary losses monétaires Knife wear tests (work done in collaboration with Vincent Blanchard) Short canter knives were coated by the ENSAM (Ecole Nationale Supérieure d'Arts et Métiers) and tested under industrial conditions in June 2010 at Abitibibowater sawmill in St. Thomas Didyme, Québec. Coated knives were identified and installed on the canter head along with new knives. The tools were tested over a 32-hour production period. The cutting edges of both coated and uncoated knives was measured before and after testing by molding the knife edge into a resin material used by dentists. A thin slice of the molding was analyzed with a microscope (usually adjusted to 20x). A digital image was captured and then measured with the AutoCAD software. Figure 6 shows a view of a canter knife edge (magnification 20x) and Figure 7 shows parameter measured.

Figure 6 Measure of a canter knife cutting edge using the AutoCAD software

127


As illustrated in the following table, knives were coated respectively with TiN, CrMoN, and CrN (2 tools). The wear of treated was 35% lower than that of uncoated tools. Table 1

Canter knife wear – Preliminary test Treatment type TiN CrMoN CrN CrN Average

Coated Wear (micron) 12.3 20.6 22.0 14.2 17.3

Uncoated Wear (micron) 14.3 38.5

26.4

Forty short canter knives were treated by the ENSAM. Ten different treatments were applied on these tools: CrN, WCCR, WCCrN, TiN, TiWC, TiWCN, MoN,CrMoN, CrVn, CrAlN. These tools have been tested under industrial conditions at the Abitibibowater sawmill in StThomas-Didyme, Québec, in February 2011, when temperatures were lower than 0 ° C. These tools, which were new before treatment, were installed along with new untreated tools on canter heads for primary log breakdown. The knives operated for a 24-hour period, which is the normal operating time during winter season. Table 2 provides the results of these tests. The average wear of the coated tools was 54.9 microns, while the average tool wear for uncoated tools was 97.1 microns; a decrease of tool wear of about 43% . The wear rate is higher than that observed in preliminary testing when the logs were at a temperature above 15 °C. This test showed that the surface treatment of tools greatly improves wear resistance, but more knives need to be tested before stating statistically the difference between coated and uncoated knvies. Best treatment for industrial conditions will also be identified at this moment. Table 2

Canter tool wear – Test batch # 1

CrN CrWC CrWCN TiN TiWC TiWCN MoN CrMoN CrVN CrAlN

Coated Wear (micron) 27.8 38.3 59.6 41.8 39.4 54.0 36.5 119.5 94.8 37.4

Uncoated Wear (micron) 75.4 75.3 48.1 30.6 168.4 185.0

Average

54.9

97.1

Treatment type

128


Plasma Reactor at FPInnovations laboratory in Québec FPInnovations in Quebec purchased a plasma reactor in December 2010. This system (manufactured by Plasmionique) is a sputter deposition reactor using three cylindrical magnetrons that can be installed at five different locations on the vacuum chamber to facilitate the treatment of all geometries. A rotating substrate polarizable holder improves the consistency and quality of deposits. This system uses inductively coupled plasma (ICP) to make plasma-enhanced chemical vapor deposition (PECVD). The dimension of the treatment chamber is 30 cm wide by 50 cm in length, which allows the treatment of a wide range of tools including 28 cm diameter circular saws. Plasma to vaporize the components in place can be created by RF (radiofrequency) discharges of 13.56 MHz with a power up to 300 W or DC (direct current), whose power can reach 500 W.

Figure 7 Deposition chamber SPT-420 (Plasmionique) by magnetron sputter deposition FPInnovations The pumping system connected to the deposition chamber consists of a primary vane pump to reach a vacuum of about 10-2 mTorr. It is connected to a secondary turbomolecular pump that ensures a high vacuum that can reach 10-7 mtorr, Circular saws have been treated with chromium nitride (CrN) deposits for secondary manufacturing plants in this study. These saws will be tested later under industrial conditions. Figure 8 shows a saw being treated with CrN.

129


Figure 8 Saw under CrN treatment CONCLUSIONS The cant shape analyser is capable to measure the geometry and surface quality of cants during lumber production. Money losses due to different problems are estimated to more than 1 million $ CDN and can be recovered by the installation of this system. Tools coating with a plasma reactor improved knife wear resistance from 35 to 43%. Further tests need to be done to identify the best tool coating treatment under industrial conditions. ACKNOWLEDGMENTS The authors would like to thank Pierre-Marc Minville, Marc Desjardins, Louis-Martin Laforge from AbitibiBowater for their cooperation and assistance during testing canter knives tests. We also wish to thank Mr. Philippe Turcot from Nap Gladu saw for the exchanges and his implication during the project. The authors are particularly grateful to Professors Corinne Nouveau (LABOMAT, Cluny, France) and Luc Stafford (University of Montreal) who treated canter knives, provided access to their processing equipment and characterization, technical resources. We also thank Martin Giroux and Martin Boyer from Lauzon for supporting this project. FPInnovations like to thank Natural Resources Canada - Canadian Forest Service for the financial support provided through the Value to Wood program.

130


REFERENCES 1.

Nouveau, C., C.Labidi, J.-P.F. Martin, R.Collet et A.Djouadi. 2007. Application of CrAlN coatings on carbide substrates in routing of MDF. Wear 263 (2007) 1291-1299.

2.

Pinheiro, D., M.T. Vieira, J.P.Dias, M.A.Djouadi et C.Nouveau. Wear delamination of mono and multilayer coatings during the cutting of wood-based products.

3.

Sheikh-Ahmad, J. and T.Morita. Tool coatings for wood machining. pp. 109-119.

4.

Beer, P. R.Marchal, J.Rudnicki, S.Miklaszewski et P.Gogolewski. Wood processing by peeling with nitrided steel knives. Processing IWMS 14. pp.507-509.

5.

Nouveau, C. R.Marchal, M-A Djouadi, M.Lambertin, G.Brun, C.Marchand and P.Beer. 1999. Deposition of hard coatings by pvd methods on cutting tools: application in wood machining. Processings IWMS- 14 , pp.441-451

6.

Laganière, B. 2011. Amélioration du procédé de débitage avec l'aide d'un analyseur d'équarri Chantiers Chibougamau, Scierie Chaleur, Abitibibowater (secteur St-Thomas), Daaquam. Four confidential reports.

7.

Laganière, B et V.Blanchard. 2011. Amélioration de la performance des outils de coupe par traitement par nanodéposition plasma. FPInnovations. 40 pp.

Contact : [email protected]

131


Detecting Top Rupture in Pinus sylvestris Sawlogs Skog, Johan1,2, Lundgren, Nils2,3 and Oja, Johan1,2 1

SP Technical Research Institute of Sweden, Wood Technology Skeria 2, SE-931 77 Skellefteå, Sweden 2 Luleå University of Technology, Division of Wood Science and Technology, Skeria 3, SE-931 87 Skellefteå, Sweden 3 Umeå University, Department of Applied Physics and Electronics, SE-901 87 Umeå, Sweden Corresponding author: Johan Skog. E-mail: [email protected]

ABSTRACT The quality of sawlogs is sometimes reduced by defects such as spike knots, cross grain and compression wood, which have been formed in the growing tree after rupture of the main stem. The hypothesis is that such damages can be found prior to sawing by using the straightness of the log as an indicator. If top ruptures can be detected already on the logs, then it is possible to treat these logs in a specific way, for instance, using lower speed or thicker blades than when sawing normal logs. This also means that, for the large volume of material, the sawing process can be optimized without problems caused by logs with top rupture. In this study, data from optical three-dimensional log scanners and X-ray log scanners were used to determine the outer shape and the heartwood/sapwood border of the logs. Indicator values based on sharpness of the crook in outer shape and heartwood shape were calculated and the logs were sorted by decreasing top rupture indicator value. The result was compared to the number of top ruptures noted in manual grading of the sawn boards. Two sets of data have been used, scanner data simulated from computed tomography images of 540 Scots pine (Pinus sylvestris L.) logs were used for algorithm development and data obtained by industrial scanning of 508 Scots pine logs were used for algorithm validation. The results show that both outer shape and heartwood shape can be used as indicator of damages from top rupture and that the heartwood shape indicator is capable of detecting logs with severe top rupture without making any false identifications.

INTRODUCTION When the top of a growing tree is broken, e.g. by moose browsing, strong winds or by a snow load, one of the branches will bend upwards and form a new leading shoot [1]. This reduces the growth and cause a deformation, especially if the main stem is broken and not just the apical leading shoot. It is also possible that two or more branches compete for several years until one of them dominates enough to form a new top. Wind and snow may cause breakage at any height while moose browsing only affects the butt log. The most common type of browsing damage is apical leader loss and is most frequent at stem heights of around 1 m. The more severe browsing damages, main stem breakage and bark stripping, are most frequent at heights of 2–3 m [2]. This means that such damages are close to the mid part of the butt log. The Swedish moose population peaked during 1981–1982 and it can be expected that this has caused a peak in browsing damages to young trees during that period [3, 4]. As the tree grows, the damage will remain inside the stem even when it has been hidden by new wood. Therefore logs with large bark strips should be removed in early thinning while the damage is still visible [5].

132


Figure 1. Board sawn from a Scots pine log where the top was broken in the growing tree.

The crook caused by top rupture also gets less sharp as the tree grows and overgrown ruptures cannot always be detected from the outer shape of the stem. The strength of centreboards sawn from such logs will however still be reduced by cross grain and compression wood [6]. Variations in density and grain direction will induce lateral forces on the saw blade [7], which put a constraint on feeding speed and minimum saw blade thickness. Furthermore, such variations will cause deformations and stress in the wood during drying and there will be more visible defects on the sawn goods, e.g. vertical, bark-ringed and decayed spike knots. Figure 1 shows a board from a log where the top was broken when the tree was young and one of the branches formed a new leading shoot. The remaining defect is more pronounced in the heartwood shape than in the outer shape of the stem, where the crook has been reduced by formation of new wood. Today, optical three-dimensional (3D) scanners are used by many sawmills to measure the outer shape of logs and can be used to calculate various variables describing the crook, such as bow height, angle and sweep [8]. Even though attempts have been made to classify logs by different types of crook [9], sweep is normally the only variable being used to describe straightness of the logs. The introduction of X-ray scanners has made it possible to include also interior properties in log sorting [10, 11]. The low moisture content in heartwood relative to sapwood gives a high contrast in the radiographic images obtained by X-ray scanning of green logs, meaning that the amount of heartwood can be measured. It has also been shown that the accuracy in measurements of the heartwood/sapwood border can be improved by combining the X-ray images with outer shape data from a 3D scanner using path length compensation. This makes it possible to obtain detailed information about the heartwood shape by means of industrial X-ray scanning [12]. The aim of this study is to develop algorithms for automatic detection of top rupture. The algorithms will be based on the outer shape and heartwood shape information available from industrial 3D and X-ray log scanners.

MATERIALS AND METHODS The Swedish Pine Stem Bank During the development of the algorithms for automatic detection of top rupture, 3D and X-ray data generated from computed tomography (CT) images was used. The algorithms used for simulation of log scanner data from CT images are described by Grundberg and Grönlund [13] and Skog [14].

133


Figure 2. X-ray image of a Scots pine log where log shape (solid), heartwood shape (solid), filtered centre of the log (dashed) and centre of the heartwood (dash-dotted) have been marked. Low density values appear dark in the figure.

The top rupture algorithms are based on observations from 540 logs in the Swedish pine stem bank [15] with top diameters in the range from 137 to 350 mm. This is a database containing detailed information about logs from 200 Scots pine trees. All logs have been scanned in a medical CT scanner (Siemens Somatom AR.T, [16]). After scanning, the logs were sawn using a normal sawing pattern and dried to 18 % moisture content. The centreboards were manually graded by a skilled grader according to the Nordic Timber grading rules [17]. About 80 % of the boards were also graded by a second grader as a reference. Visible defects on the boards were noted by both graders, but normally only the worst defect was noted if there was more than one kind of defects in a board. Because the grading did not explicitly focus on defects caused by top rupture, the number of boards noted as containing such defects may be too low. Data For Industrial Validation Of Algorithms The algorithms were validated by a second set of data obtained by industrial scanning of 508 Scots pine logs with top diameters in the range from 160 to 300 mm. These logs were scanned using a two-directional RemaLog XRay scanner [18] in line with a RemaLog Bark 3D scanner [18] in the log sorting station of a Swedish sawmill. 39 of the logs were manually selected because their outer shape revealed that they were damaged by top rupture and another 469 logs were taken from the normal production. The sawn boards were manually graded in the green sorting and top ruptures were noted. Detection Of Top Rupture Using Heartwood Shape 3D and X-ray log scanner data are combined using path length compensation. For each crosssection of the log, the 3D scanner measures the outer shape of the log and the X-ray scanner measures the attenuation of X-rays passing through the cross-section. The 3D cross-section is then used to calculate the path lengths travelled through wood for X-rays arriving at each detector pixel. By combining the measured attenuations and the calculated path lengths, the average green density along each individual X-ray path is calculated. The final result is a green density image of the whole log with good contrast between heartwood and sapwood [12]. Because a twodirectional X-ray scanner is being used, two perpendicular images are obtained for each log. The outer contour and heartwood contour of the logs are obtained from the green density images [12] and a short smoothing filter is applied to the contours. Log and heartwood centre lines are calculated from the detected contours and a long averaging filter is used to even out variations in the log centre line (Figure 2). A signal containing local heartwood shape irregularities is then extracted by subtraction of the filtered log centre line from the heartwood centre line. For each point in the irregularity signal, the gradient of the signal is calculated over a window to the left and to the right of the point respectively. In positions where the two gradients have opposite

134


signs, their absolute values are summed and used to indicate a sharp deviation in the shape. Finally, a filter is applied to reduce indications close to the ends of the log because of the higher uncertainty in measured shape at the log ends. Because a two-directional X-ray log scanner is being used, this process is repeated for both measurement directions and the maximum irregularity indicator value found in any of these two directions is being used as a top rupture indicator of the log. Top Rupture Detection Using Outer Shape Only Two different indicators for detection of top rupture using outer shape only are also evaluated. In the first approach, outer shape from the X-ray scanner is being used. The algorithm for heartwood shape irregularities described above is being used, but applied to the outer shape instead of the heartwood shape. In this case, the unfiltered log centre line is used in place of the heartwood centre line, meaning that the irregularity signal will describe local deviations in the outer shape rather than in the heartwood shape. In the second approach, the outer shape measured by the 3D scanner is being used. Still, the same algorithm is being applied. In this case, the 3D shape is simply projected onto a plane along the main axis of the log, yielding an outer contour similar to the one in Figure 2. The centre line of this contour is then used in place of the log centre line measured by the X-ray scanner and the rest of the calculation proceeds as described above. In order to find the worst shape irregularities, the process is repeated eight times using different projections of the 3D shape. The planes being used for the projection are rotated 0, 22.5, 45, 67.5, 90, 112.5, 135 and 157.5 degrees around the main axis of the log respectively. The maximum irregularity indicator value found in any of these eight directions is being used as the top rupture indicator of the log. Evaluation Of Results Logs where the graders had found top rupture on any of the sawn boards are considered being top rupture logs. Each of the three top rupture indicator values, based on deviations in the heartwood shape, outer shape measured by X-ray and outer shape measured by 3D respectively, are then graphically compared to the top rupture notation (yes/no) from the graders. Suitable threshold values for finding logs with top rupture are chosen for all three indicators and logs with high indicator values but no noted top ruptures are examined more closely. The different top rupture indicators are also pair wise compared to each other in order to find any differences in which logs the methods tend to find.

RESULTS The heartwood-shape and outer-shape top rupture indicator values of the logs from the Swedish pine stem bank are plotted in Figure 3. For all three methods, it was found that the highest values correctly correspond to logs where the sawn boards had been graded as top rupture boards. In order to facilitate comparison between the methods, each series was normalized so that the highest indicator value of any log not marked with top rupture by the board graders corresponds to a value of 10. For each method, the 15 logs with the highest indicator values that had not been graded as top rupture logs were investigated further. This was done by inspection of the CT images from the stem bank. It was found that two logs (squares in Figure 3) were severely damaged by scars, over a large portion of the log the sapwood was completely missing. This type of logs is not suitable for sawing and should thus never appear at a sawmill, therefore these two logs were excluded from further analysis. The CT image inspection also showed that, for each of the three methods, at least the top 10 logs that had not been graded as top rupture logs actually did contain top rupture (diamonds in Figure 3).

135


Figure 3. Top rupture in 540 Scots pine logs from the Swedish pine stem bank. The graphs show manual top rupture notation versus automatic top rupture indicator values based on heartwood shape from X-ray and 3D log scanner, outer shape from X-ray log scanner and outer shape from 3D log scanner respectively. Circles are top ruptures noted during board grading, diamonds are top ruptures found when investigating computed tomography (CT) images of the logs and squares are logs with big scars. The dashed lines are suggested threshold values for separation between logs with and without top rupture. The calculations are based on X-ray and 3D log scanner data simulated from CT-scanned logs.

Figure 4. Top rupture in 508 Scots pine logs from a Swedish sawmill. The graphs show manual top rupture notation versus automatic top rupture indicator values based on heartwood shape from X-ray and 3D log scanner, outer shape from X-ray log scanner and outer shape from 3D log scanner respectively. Triangles are logs presorted as top rupture logs and circles are top ruptures noted during board grading. The dashed lines are suggested threshold values for separation between logs with and without top rupture. The calculations are based on industrial X-ray and 3D log scanner data.

The value 10 was chosen as a threshold for separation between logs with and without top rupture. This value is the highest value of a log not noted with top rupture during board grading, and the CT image inspection revealed that all logs with values close to 10 did contain top rupture. Therefore, this choice of threshold value left a margin between the threshold and the first logs without top rupture, see Figure 3. The feasibility of this threshold was then evaluated by comparison with the corresponding industrial measurements, see Figure 4. Table 1 shows the number of logs being correctly and incorrectly sorted as top rupture logs when using different threshold values.

136


Table 1. Number of logs being correctly and incorrectly sorted as having top rupture (TR) when applying different threshold values to the top rupture indicators obtained using the algorithms based on heartwood shape (Heart), outer shape from X-ray (OuterX) and outer shape from 3D (Outer3D). The total number of logs in the stem bank data set is 540, whereof 66 with top rupture (including the top ruptures from grading and from inspection of CT images). The total number of logs in the industrial data set is 508, whereof 76 with top rupture. Stem bank data Algorithm Threshold Correct TR Incorrect TR Heart 8 9 0 Heart 10 6 0 Heart 12 4 0 OuterX 8 8 0 OuterX 10 4 0 OuterX 12 1 0 Outer3D 8 4 0 Outer3D 10 4 0 Outer3D 12 2 0

Industrial data Correct TR Incorrect TR 27 7 19 3 11 1 9 2 5 2 3 1 23 25 18 9 14 5

Figure 5. Automatic top rupture indicator values for 540 Scots pine logs from the Swedish pine stem bank. The left figure shows indicator values based on heartwood shape from Xray and 3D log scanners, versus indicator values based on outer shape from an X-ray log scanner. The right figure shows indicator values based on outer shape from a 3D log scanner, versus indicator values based on outer shape from an X-ray log scanner. Circles are top ruptures (TR) noted during board grading, diamonds are top ruptures found when investigating computed tomography (CT) images of the logs, squares are logs with big scars and crosses are logs without top rupture. The calculations are based on X-ray and 3D log scanner data simulated from CT-scanned logs.

137


Figure 6. Automatic top rupture indicator values for 508 Scots pine logs from a Swedish sawmill. The left figure shows indicator values based on heartwood shape from X-ray and 3D log scanners, versus indicator values based on outer shape from an X-ray log scanner. The right figure shows indicator values based on outer shape from a 3D log scanner, versus indicator values based on outer shape from an X-ray log scanner. Triangles are top ruptures found in log sorting, circles are top ruptures found on the boards in green sorting and crosses are logs without top rupture (TR). The calculations are based on industrial X-ray and 3D log scanner data.

In Figure 4 it can be seen that many of the logs that were presorted as containing top rupture get high indicator values from both the heartwood indicator and the outer shape indicators. However, a number of logs not containing any top rupture also get high values from the two outer shape indicators, making it difficult to use the outer shape methods to separate between logs with and without top rupture. The relationship between the different top rupture indicators is being shown in Figure 5 for stem bank data and in Figure 6 for industrial data.

DISCUSSION Images that are produced by an X-ray log scanner contain a lot of information and a requisite for successful application of the technique is the ability to extract and handle the most useful parameters. The heartwood shape is one such parameter and the results show that it can be used to detect top ruptures that cause crooking of the heartwood. For the stem bank data, the heartwood method as well as the methods based on outer shape, measured either by the X-ray scanner or the 3D scanner, were all capable of detecting a fair amount of the logs with top rupture without giving false alarms. The highest indicator values corresponded to top ruptures noted in the grading and inspection of the CT images of the logs with high indicator values but no notes about top rupture on the boards showed that most of them contained top rupture. The reason that these top ruptures had not been noted could be that some of these boards had other defects that were being noted instead or, more likely, that the top ruptures were not visible or hardly visible on the boards. This means that the top rupture indicators also give information on the severity of the top rupture, and by tuning the threshold value, it will be possible to identify only logs with severe top rupture or also logs with less severe top rupture.

138


The industrial validation of the algorithms shows that the heartwood shape algorithm works fine also on industrial data. The algorithms based on outer shape did however not behave very well in the industrial validation because some logs without top rupture gives rise to high values (Figure 6). Especially the algorithm based on outer shape from the 3D scanner proves useless as the highest values are caused by logs without top rupture. The reason for this could be that the logs have been turning or bouncing around on the conveyor as they were passing through the 3D scanner. The only other log defect found in this study that can cause a strong top rupture signal in the heartwood shape algorithm is the scars. However, because these scars are so big (up to one third of the log cross-section is missing), these logs would normally have been sent directly to a pulp mill or sorted out as soon as they arrive at the log sorting of a sawmill. Smaller scars could also be expected to give a signal in the heartwood shape irregularity algorithm; but because scars usually only affect one side of the log, the centre line will not be strongly affected and the signal from a scar should be expected to be weaker than that of a top rupture with a well-defined heartwood crook. The conclusion is that the algorithm based on irregularities in the heartwood shape should be a good candidate for the automatic detection of top rupture, with very low risk of making false positives. Implementing top rupture identification in a sawmill would be a great tool for avoiding the problem logs when making products that are sensitive to top rupture. It will however not be desirable to completely remove the detected logs from production, instead, the top rupture logs could be used for alternative products that are not sensitive to top rupture, one example being wood for finger-jointing. Because of the rapid variations in density and grain direction near the top rupture, it is one of the factors limiting feeding speed and saw blade thickness. Therefore, an algorithm for automatic detection of top rupture can be expected to play an important role in a system optimizing the sawing to the raw material. An important task for the future will be the development and implementation of such strategies.

ACKNOWLEDGEMENTS This work was financially supported by TräCentrum Norr, a research programme jointly funded by industrial stakeholders, the European Regional Development Fund (ERDF) and the county administrative boards of Norrbotten and Västerbotten.

REFERENCES 1. Mattheck, C. & Kubler, H. (1995) Wood – The Internal Optimization of Trees. Berlin: Springer-Verlag. ISBN 3-540-59318-7. 2. Bergqvist, G., Bergström, R., & Edenius, L. (2001) Patterns of stem damage by moose (Alces alces) in young Pinus sylvestris stands in Sweden. Scandinavian Journal of Forest Research, 16, pp 363–370. 3. Bergström, R. and Vikberg M. (1992) Winter browsing on pine and birch in relation to moose population density. Alces supplement 1, 1992, 127–131. ISSN 0835-5851.

139


4. Hörnberg, S. (2000) Changes in population density of moose (Alces alces) and damage to forests in Sweden. Forest Ecology and Management, 149, 141–151. 5. Karlmats, U., & Pettersson, N. (2001) The effect of moose grazing on wood properties in Scots pine. Högskolan Dalarna. Rapport 12b. ISSN 1403-8188. (In Swedish with English abstract). 6. Timell T.E. (1986) Compression Wood in Gynosperms vol.3. Chapter 15. Berlin: Springer-Verlag. ISBN 3-540-15715-8. 7. Axelsson B (1993) Cutting Forces at a Tool Edge during Machining of Wood. Licentiate thesis 1993:22L. Luleå University of Technology, ISSN 0280-8242. 8. Lundgren, C. (2000) Predicting log type and knot size category using external log shape data from a 3D log scanner. Scandinavian Journal of Forest Research, 15, 119–126. 9. Gjerdrum, P., Warensjö, M., & Nylinder, M. (2001) Classification of crook types for unbarked Norway spruce sawlogs by means of a 3D log scanner. Holz als Roh- und Werkstoff, 59, 374–379. 10. Aune, J. (1995) An X-ray Log-Scanner for sawmills. In Lindgren, O. (ed.). Proceedings from the 2nd international seminar/workshop on scanning technology and image processing on wood. Technical Report 22 T, pp. 51-64. Luleå University of Technology. ISSN 0349-3571. 11. Grundberg, S. & Grönlund, A. (1995) The development of a LogScanner for Scots pine. In Lindgren, O. (ed.). Proceedings from the 2nd international seminar/workshop on scanning technology and image processing on wood. Technical Report 22 T, pp. 39-50. Luleå University of Technology. ISSN 0349-3571. 12. Skog, J. & Oja, J. (2009) Heartwood diameter measurements in Pinus sylvestris sawlogs combining X-ray and three-dimensional scanning. Scandinavian Journal of Forest Research, 24, 182–188. 13. Grundberg, S. & Grönlund, A. (1997). Simulated grading of logs with an X-ray LogScanner – grading accuracy compared with manual grading. Scandinavian Journal of Forest Research, 12, 70–76. 14. Skog, J. 2009. Combining X-ray and 3D scanning of Logs. Licentiate thesis, Luleå University of Technology, Sweden, 26 pp. ISSN: 1402-1757. 15. Grönlund, A., Björklund, L., Grundberg, S. & Berggren, G. (1995) Manual för furustambank [Users manual for the pine stem bank]. Luleå Univ. Technol., Teknisk rapport 19 T, 25 pp. ISSN 0349-3571. (In Swedish.) 16. Anon. (2011) Siemens Healthcare, Erlangen, Germany. http://www.medical.siemens.com/. Accessed March 30, 2011. 17. Anon. (1994) Nordic Timber, Grading rules for pine and spruce sawn timber (The blue book). Stockholm: The Association of Swedish Sawmillmen. ISBN 91-7322-227-5. 18. Anon. (2011) RemaControl Sweden AB, Västerås, Sweden. http://www.remacontrol.se/. Accessed March 30, 2011.

140


Sensory Quality Assessment Of Surfaces, Especially Wooden Surfaces Riegel, Adrian; Dekomien, Kerstin and Baade, Jan Christian University of Applied Sciences Ostwestfalen-Lippe, Lemgo, GERMANY

ABSTRACT The quality assessment of wood surfaces is insufficient in its current state. Nowadays measurements are possible but the technique and the results require a lot of experience. They are also too small in range and too expensive to meet the requirements in practice. Traditional methods of assessment are based on experience and intuition without any standardisation. The paper presents the current status of sensory testing methods whose ambitions are to make the assessments more comparable and should be standardised in the VDI guideline 3414-2. This guideline will respect both ways for quality assessment, measuring and sensory testing. The implementation of these methods opens up the opportunity to determine a specific quality level and improve it ongoing. Intensive instruction enables the reduction of the number of testing personnel with only small significant differences in the quality level.

INTRODUCTION Wooden surfaces, especially the ones of furniture, are mainly judged from the customer by their visual and haptic character. In everyday practice this judgment is based on intuitive personal methods. The communication of the correlated experiences is mainly insufficient and the test procedure is in consequence often carried out only by few specialists. Results gathered this way are hardly comparable and reproducible. On the other hand common measurement methods and the linked characteristic values, like the profile method with stylus instruments, are in the field of woodworking not appropriate for a comprehensive industrial usage, since they are mainly designed for functional and not aesthetic surfaces, often expensive and have only a small measurement range, which reflect only a small image of the surface. Making the imprecise personal quality inspections reproducible and comparable by introducing scientific sensory methods is the ambition of a research project carried out at the University of Applied Sciences Ostwestfalen-Lippe. The food technology uses sensory quality assessments effectively for years. Analogous to those methods guidelines for selection and instruction for the assessment personnel as well as guidelines for visual and haptic quality assessments are developed. Therefore common methods are picked up and aligned to the requirements of the wood and furniture industry. Sensory methods are not to replace the metrology in fact they are an addition and correlations are welcome.

SENSORY ASSESSMENTS IN FOOD INDUSTRY AS MASTER SAMPLES In the food industry exist a large number of specific national, European and international standards, e.g. ISO 4120 (Triangle Test) [1]. Both industry food and wood industry deal with similar products made out of natural resources, being picked by the customer due to sensuous affectation, with a high range of different feature presentations. The decision purchasing a product is rather based on intuition than on rational and measureable reasons. Some attributes are

141


measureable e.g. the concentration of salt in a product, but this aspect alone is not suitable for an evaluation since the overall impression the interaction of various attributes- turns the balance and generates the flavour. The same can be said for wooden surfaces. A surface evaluation can hardly be broken down on one single attribute and its characteristics. The customer examines the whole visual and also haptic impression of the product, therefore the sum of attributes and their obvious existence leads the customer to his decision. These requirements are similar to the ones in the food industry. The used senses might be different, but the methods themselves are established for decades. In the wood industry instead of olfactory and gustatory, haptic and visual impressions are the most important ones. But the problems are the same. On the one hand expects the customer a high quality, but mostly he is not able to describe or define the expected quality as well as the manufacturer itself [2]. On the other hand needs the manufacturer reliable, reproducible and standardised methods for evaluating and determining the said quality. Expecting that measuring leads to one characteristic value describing this “unsharp” overall quality will fail due to the complexity of the quality description, the evaluation processes related to the expectations of the customers and its disturbances. Therefore it is necessary to define exactly these variations in quality level, disturbances or flaws into cluster also according to their location of appearance in the process.

EXAMPLE HIGH GLOSS SURFACES High Gloss surfaces require if lacquered a very complex layer construction, which influences the surface quality [3]. Moreover building up pervasive cycles for quality control needs a two dimensional view. On one hand the expectations of the customers should be satisfied, however fuzzy the attributes are. On the other hand the quality evaluations should in case of flaws finally help to eliminate their sources in the process chain. Defining and evaluating all relevant disturbances in this sense requires a wide knowledge of optophysics, lacquering process and lacquering systems. This diversity of required knowledge explains the difficulties in defining relevant factors. Figure 1 shows a simple model of a lacquering process chain, beginning with the sanding process and ending with UV-drying the top layer. In the shown process are a lot of possible positions where it is useful to evaluate the quality, according to their point of occurrence.

Figure 1: Processchain Lacquering Wooden Surfaces Measuring all parts e.g. at point 1 with the profile method is not practicable, since the feed rate for spraying is up to 8 m/min and the output is not just one piece. The profile method gives just data for the profilometry and not for the whole impression. Since wood is an inhomogeneous

142


material, it is difficult to measure and interpret the results. It is not possible to conclude from the profilometry to the exact surface impression. Both, the profile method with a stylus instrument and the haptic or visual evaluation, cannot be done in an online mode. Samples have to be taken throughout the whole process in order to be evaluated [4]. Measuring devices generate specific values concerning one characteristic, but they are usually not able to reflect the whole surface impression. The rules and procedures for the stylus method, described in ISO 4288 [5], specify the location on the surface where to measure. The profile has to be taken at the expected worst case location, selected and evaluated by a human. This selection is based on intuition and experience, but this evaluation is never holistic and focused on only one defect or attribute. In order to give a correct impression of the surfaces it is necessary to have a sensor, which is capable to measure multiple values e.g. roughness, fibrousness or uniformity of the sanding marks, in a short period of time with a onetime measurement.

PROCEDURE TO ACHIEVE THE GOAL Due to the discrepancy of a single attribute measured by experts and the holistic impression of finished surfaces by customers, the research project of sensory quality assessment was born. In order to embrace the requirements of the customer it is essential to define the characteristics, their attributes and the specification. This classification is necessary because one characteristic can be caused by various factors. The roughness for example can be reflected by the sanding material, the differences in the early- and latewood or the fibrousnesses. Are these attributes defined, they can be evaluated according to their specification. Mostly the intensity, the absence or the existence of an attribute is the factor to be evaluated. For teaching and testing purposes different surface characteristics were analyzed and described with attributes.

Figure 2: The Method of Sensory Quality Assessment After identification of all relevant standards they were tested in various experiments in order to proof their capability on wooden surfaces. Therefore different assessor panels were set up and taught these methods as shown in Figure 2. The preselection of the naive assessors was based on the motivation and interest for this way of quality assessment. Furthermore the visual and tactile constitution of each naive assessor was tested. For training and afterwards testing purposes

143


different samples, containing the attributes with differences in substrates, characteristics and their specifications, were created and tested by the panels. After training the qualified assessors were ascertained by statistical methods. To make sure that the conclusion quality of the panel remains good enough, the panels had to pass training time by time. Sensory Assessor Preselection The selection of assessors is one important aspect in the method. Possible assessors must fulfil certain requirements in order to become a full assessor. Criteria are elementary divided in physiological and psychological aspects in order to verify sufficient haptic and visual recognition thresholds, the motivation or mental attitude concerning their task, like it is shown in XY-Theory [6]. The physical selection includes an eyesight test and a colour vision test describing the visual recognition of the naive assessor. Haptic recognition thresholds are hardly identified by common and simple medical tests only the two point test gives an evidence for level of perception. The haptic–research laboratory at the Leipzig University is actual working on a haptic–recognition threshold test for using it in different medical and industrial purposes. This test gives an informative and comparable result about the tactile recognition [7]. According to this test a prediction concerning the haptic resolution seems to be possible, so this result can be used to select naïve assessors. In order to compare different panels in the start, a panel consisting out of blind persons was used as a reference group for standard tests, since they were more sensibiliser than the other groups due to their capability of reading braille. Sensory Assessor Training After the selection of each naive assessor the training started with the development of a consistent vocabulary for all attributes. Therefore it was necessary to divide these attributes of sanded surfaces in geometric, optical, structure and chemical-physical and subdivided in regular and irregular properties [4]. Another aim of the training is to increase the perception. Therefore coated abrasive products and sanded PMMA samples were used due to their simple differentiation and rigidity against abrasion during tactile testing. Besides this permanent training the panels were taught in different testing methods analogous to food industry. The procedure includes for example triangle tests with coated abrasive products with a grain of P600 and P800 under the provisions of FEPA [8], which means a difference in size of grain of approx. 4 µm. Furthermore they had to find out these sanded differences on pieces of polymethylmethacrylate (PMMA). All panels got the same basic information about the tactile perception and different movements for fingertips or hand and how to get an overview to all examples. The further focus of sensory assessor training was getting in touch with a diversity of surfaces and substrates in order to study several attributes of sanded surfaces. Because of the fast visual information processing and for an increased recognition of individual wood sanded attributes or characteristics, the panels had to test the surfaces tactile before they were allowed to check their decision visually. The way of testing tactile and visual the naive assessor had the possibility to learn each attribute with two sensations, which links their knowledge better. When testing tactile the perception is successively and takes more time to recognise characteristics of objects. At first there is only a global understanding while learning tactile, further steps increase the recognition of characteristics of surfaces, especially for classification [9]. Motivation of the assessors is one of the most important aspects, without it assessors or a whole panel cannot retain their recognition level. The motivation is influenced by many different facts

144


like illness, stress or time of working rather testing, as well as knowing the importance of surface quality assurance or following consequences because of defects. Different approaches about motivation indicate these facts. By an examination at the end of training it could be achieved to become or to stay an assessor. Statistical Sensory Assessor Selection Performing successful assessments requires a selection of sufficient trained assessors, concerning all relevant attributes. With the sequential analysis of Wald the learning success is documented and gives an impression for the deficit on different attributes for each naive assessor [10]. Advantage of this analysis is the smaller sample size needed and the good visibility of the results in a simple graphic way. According to Wald three areas are defined by statistical methods. One area of refusal, one of further teaching and one of accepting. Based on the analysis of naive assessors, the assessors with a remote recognition (area one) while they passed the preselection, should not stay in the panel. Naive assessors with a less deficit or a problem with only one attribute have to pass the training again (area two). Assessors with a sufficient recognition (area three) can be instated as assessors. Regular Sensory Assessment In the research project the assessors were subdivided into four different panels with different backgrounds concerning their knowledge of the raw material and its working processes. The first trained panel consisted out of six blind persons. The second panel were ten students, studying wood engineering. Trained students from a wood department of another college were the third panel. In order to compare the results and achievements, 24 students studying wood engineering were referred as control group without any further teaching or training (fourth panel). The panels performed a variety of tests according to the international standard. The basic methodology for a typical difference test, the triangle test is shown in Figure 3. Grading tests were also carried out, with similar results.

Figure 3: Schematic Flow Chart of the Triangle Test The following results refer only to haptic perception and triangle tests. Figure 4 shows the perception level of these panels performing a test on sanded MDF. A triangle test consists out of three probes. Two of these probes (e.g. P120) are in a similar way prepared and the third one is prepared a different way (e.g. P180, or vice reverse). The overall perception level did not exceed 62%. Examining the five best persons, from the trained students and trained college students,

145


these results are shown in Figure 5, their overall perception level is close to the 80% mark. Differentiating these persons in first mentioned triangle test, shows a perception level of 100% for this outstanding assessors. These results show the necessity of a severe assessor selection. All the participants passed the preselection. But as Figure 1 shows, most of them were not able to transfer their haptic knowledge on to wooden surfaces. The second last column in Figure 5 shows the results of a triangle test on an MDF substrate sanded with P220 aluminium oxide abrasive alternatively silicon carbide. This haptic test seemed to be difficult for the assessors. Looking at the group of the best assessors, a perception level of 20% has been achieved. A visual evaluation of the same probes showed a perception level of nearly 90%. Therefore it seems to be necessary to combine the visual and haptic evaluation of wooden surfaces during regular sensory assessment.

Figure 4: Perception Level of Differing Panels

Figure 5: Outstanding Assesors Table 1 shows some surface characteristic values determined by the profile method. Concerning these results, there is just a slight difference. Nearly all shown characteristics are in the range of dispersion of characteristic belonging to the other surface. The differences are haptic not detectable, but visual there is still a difference noticeable (90% of the assessors noticed the difference). A combination of the haptic and visual assessment would be solving the problem. The strength of the research project is the structured methodology in the testing. A structured and defined way for assessing surfaces is one opportunity to minimize errors and differences in the individual results of each assessor. Figure 3 shows the methodology for the triangle test. The main steps are similar for all tests. Defining a clear, unmistakeable objective is the first step for a successful assessment. Sensory assessments have to be done in a defined environment. Changing conditions e.g. in the luminous colour or luminance leads to deferring results, as tests also carried out in the research project showed. In order to eliminate these disturbances it is useful to install

146


some tools, which support the assessors during their assessments. These tools can be special lights, arranged in a way so they give a sharp, directed beam over the surface. Such a beam was developed [11]. Table 1: Comparison of Different Surface Characteristics Using Profile Method Measurement conditions:

Characteristic [µm] Ra standard deviation Rz standard deviation R standard deviation Rx standard deviation Rpk standard deviation Rvk standard deviation

lt: 15mm n: 5

lc: 2,5mm

A: 500µm B: 2500µm Vt: 0,5mm/s Grinding: P220 (FEPA) Unit: µm Abrasive Aluminium Oxide

Filter: ISO 11562 Silicon Carbide

1,975

1,8

standard deviation compared to characteristic

0,126

6,37%


18,525 2,134 11,52%

0,122 1,522

8,2 standard deviation compared to characteristic

0,516

6,30%

0,451

9,44%

23,52 5,926 25,20%


2,218


2,225 0,419 18,85%

5,60%

1,68 0,421

4,525 0,435

8,02% 8,04

23,5


6,80% 18,98

25,04% 5,46

9,61%

0,666

12,19%

Another possible tool is a piece of cellophane, like the ones from a cigarette box, pulled over the finger. The cellophane functions like a mechanical filter on the fingers which suspends characteristics with a short wavelength. This can be necessary if the relevant characteristic is heterodyned by other ones with a shorter wavelength. Furthermore should disturbances like noise, dust or passing colleagues be ruled out by creating a defined assessment room, with restricted access during assessments. All these already developed tools will now be applied in further tests to develop a complete method for sensory assessment of wooden surfaces.

DISCUSSION In order to achieve comparable, reliable and reproducible results, some essential procedures are necessary, as shown in Figure 2. A careful and strict assessor selection is the first step achieving reliable results. Even all assessors passed the assessor preselection and training, not all were able to adapt their knowledge to different wooden substrates. Especially the panel consisting of blind persons with high tactile sensitivity had severe difficulties evaluating wooden surfaces. The main problem has been the complexity of the natural substrate wood. Characteristics, caused by the preparation process like sanding marks are interfering with characteristics coming from structure of the wood like the porosity or the difference between early- and latewood. These aspects had to be explained to panel and taught very detailed, so they gain the ability to perform assessments on wooden substrates. The knowledge of natural structure of the substrate is a necessity for each assessor and should be integrated in the assessor selection. The sensory memory is short termed and its capacity is limited. According to experiences gained during assessments and the time between different assessments, it can be said that the memory is not able to remember attributes and their specific profile, for more than one month [12]. Therefore should just the most common characteristics be taught and regularly tested and refreshed as well as discussed with the becoming assessors. Otherwise different names for one characteristic lead to confusion. A catalogue containing all relevant and possible characteristics, if possible with pictures, definition and description seems to be useful for the assessors.

147


CONCLUSION In summary it can be said that, sensory assessments of wooden surfaces are possible, under certain restrictions. The VDI guideline 3414, presented at the 19th International Wood Machining Seminar in Nanjing [4] is actually reviewed in order to account also the sensory methods. The guideline will be subdivided into four parts. Part one contains all basic rules and attributes for the quality assessment of wooden surfaces. Part two contains measuring methods, sensory methods as well as the traditional methods. Part three specialises the milled, drilled, turned and sawn surfaces. Part 4 specifies sanded surfaces. To achieve the goal of a satisfying quality assessment of wooden surfaces a consensus understanding of quality attributes and related measuring and sensory testing methods must be developed. In some cases measurements will be more accurate or reproducible to the specific attributes in others sensory testing. The research team at the University of Applied Science OWL is looking forward to develop a method to combine these both unequal methods to one overall quality concept for evaluation.

REFERENCES 1.

ISO 4120 (2004) Sonsory Analysis - Methodology - Triangle Test.

2.

Beyer L., Grunwald, M., (2001) Der bewegte Sinn: Grundlagen und Anwendungen zur haptischen Wahrnehmung. Birkhäuser, Switzerland.

3.

Dekomien, K., Kortüm, C., Riegel, A. (2011) Qualitätskriterien und Bewertungsmethoden für Hochglanzoberflächen. Holztechnologie, 52 (2): 38-43.

4.

Großmann, M., Ratnasingam, J., Rehm, K., Riegel, A., Scholz, F.(2009) Assessment of sanded surfaces according to VDI guideline 3414. Pp. 323-332. In: Zhou, H., Zhu, N., Ding, T. (Ed) Proceedings of the 19th International Wood Machining Seminar 2009, Nanjing.

5.

ISO 4288 (1996) Geometrical Product Specifications (GPS) -- Surface texture: Profile method -- Rules and procedures for the assessment of surface texture.

6.

Hohlbaum, A., Olesch, G., (2008) Human Resources, Modernes Personalwesen. Merkur, Germany.

7.

Grunwald, M. Haptic-Labor Universität Leipzig. Haptische Schwellenwertermittlung. Telefongespräch, 15.02.2011.

8.

Fachverband Elektrokorund- und Siliziumkarbid- Hersteller e.V. und Verein deutscher Schleifmittelhersteller e.V. (1984) FEPA P 43-D-1984 R 1993 - Körnungen für Elektrokorund und Siliciumcarbid für Schleifmittel auf Unterlagen.

9.

Grunwald,M., (1998) Haptische Reizverarbeitung und EEG-Veränderungen. Dissertation Universität Jena.

10.

Wald, A. (1966) Sequential Analysis. Wiley, USA.

11.

Küster, K., Riegel, A. (2008) Fotografische Dokumentation von bearbeiteten Holzoberflächen. Holztechnologie, 49 (5): 23-26.

12.

Grunwald, M.,(2008) Human haptic perception: basics and applications, Birkhäuser, Switzerland.

148


Monitoring of Measurement Accuracy - an Industrial Case David Englund SCA Timber AB – Rundvik Sawmill Box 3, 914 29 Rundvik, Sweden Corresponding author: [email protected] +46(0)70 – 670 66 23

ABSTRACT As the material costs in Swedish sawmills are about 70% and the largest parts of the total costs when making wood products the industry strives not only to get the highest outcome from the raw material but also be able to trace a product through the whole process. One way is to create accurate and significant quality measurements reliable enough to see effects of for example new maintenance routines, technical improvements or just detect errors in the process. In this investigation the thickness of the centre planks were measured with a laser camera technique just after the gang saw. The camera equipment was used to follow up how different parameters affected the sawing variations and how well the wanted green target sizes were met. The studies included several months of full industrial production. With collected data and monitoring from live production measured on the boards after the rotary gang a reference is created. This shows the effects of keeping accuracy on the saw machine setup to get out more of the raw material and further on make a better product for the planar mill. With tighter tolerances on the guides of the saw blades, regular maintaining routines and for example a consciously choice of blades the same product can be made with a tighter green target size and with yearly production of about 250 000m3 save millions. The method do not only have effect on the chosen products green target size it also prevents unexpected stops and total failures and breakdowns of the production. The approach can also be applied on other machine groups and motivate quality thinking though the process and emphasize the importance of good maintaining.

149


INTRODUCTION In Swedish sawmills about 70 % of the total costs can be derived from raw material costs. Hence it is very important for sawmills to obtain highest possible yield. The green target size is one important parameter that affects the yield and it is therefore important to optimise the green target sizes for every sawing pattern and for every piece of lumber. A reduction of the green target size at Rundvik sawmill with 0,5 mm will result in increased income with about 5 million SEC. The needed green target size is dependant on where in the cross-section a piece of lumber is cut and how large the sawing variations are [1, 2]. The sawing allowance has to be roughly 1,5 – 2 times the standard deviation of the sawn board thickness. In order to obtain a proper control the sawing variations, continuously measurements of the sizes are needed. This is especially important when different technical changes such as thinner saw blade, or changes in the saw guide design e.g. are performed. The objective of this paper is to show how we at SCA Timber, Rundvik sawmill have used a camera equipment to follow up how different parameters affect the sawing variations and how well the wanted green target sizes were met.

MATERIAL AND METHODS The study was carried out at SCA Timber Rundvik sawmill. The annual production of the mill is about 250 000 m3 sawn spruce (Picea abies). The mill is using the cant sawing methodology on an ARI profiling sawline, a sketch of the ARI gang saw is shown in Figure 1.

Figure 1. Sketch of the gang saw. In this investigation the thickness of the centre planks were measured with a laser camera technique just after the gang saw, Figure 2. The camera equipment was used to follow up how different parameters affected the sawing variations and how well the wanted green target sizes were met. The studies included several months of full industrial production.

150


Figure 2. Schematic layout of the thickness measurement equipment

RESULTS Figure 3 shows the boards from left to the right with the first piece from the left V2, V1, H1, H2 looking in the same direction as the wood comes in to the blades. The target thickness is 23,3 mm. The figures in these diagrams are measured only one the top side of the board according to Figure 2.

Figure 3 - Board split up sketch All tests were done during normal production and were performed without making any extra stops. An example of how the thickness measurements can be used is shown in Figure 4 to Figure 6. These Figures show how the thickness variations are changing during sawing of 7602 logs. The within board standard deviation is increasing with production time, in this case especially for board V2 and to some extent also for board H1 and H2. Wear of the tooth edges is a very probable cause to this increase of the sawing variations. It is also evident that the mean thickness for some of the boards is changing during the production time. Look for instance how the mean value of board HI is increasing while the mean value of board H2 is decreasing.

151


100713_624973_153_21x95_4X_2310

Number of logs

500 400 V2

300

V1

200

H1

100

H2

0 -100 22

23

24

25

26

Board width (mm)

Figure 4 - Board Thickness- TEST 1 - The first 2310 logs. 100713_624974_143_21x95_4x_723

Number of logs

200 150

V2

100

V1

50

H1 H2

0 -50

22

23

24

25

26

Board width (mm)

Figure 5 - Board Thickness- TEST 1 - The second 723 logs.

100714_624945_141_21x95_4X_4569

Number of logs

1000 800 V2

600

V1

400

H1

200

H2

0 -200 22

23

24

25

26

Board width (mm)

Figure 6 - Board Thickness- TEST 1 - The last 4569 logs.

152


The graphs in Figure 4 to 6 show how the scanning equipment can be used to analyze the measurement and the measurement variations on a time perspective of roughly one day. The data from the equipment can of course also be used for long term analyzes as shown in the example in Figure 7. We have also other examples how the scanning equipment in the daily continuously improvement work. Some of these examples will be presented at the oral presentation.

Månadens standardavvikelse 21x95

0,93 1,00 0,90 0,76 0,80 0,70 0,66 0,660,65 0,70 0,620,580,63 0,580,61 0,54 0,60 0,52 0,49 0,47 0,50 0,42 0,40 0,35 0,40 0,32 0,30 0,20 0,10 0,00 T3

T2

T1

Apr

Mar

Feb

Jan

Dec

Nov

Okt

Sep

Aug

Jul

Jun

Maj

Apr

Mar

Feb

Jan

Figure 7. Monthly standard deviation for the product 21 x 95 mm.

DISCUSSION AND CONCLUSIONS Our experience from our daily process improvement work at Rundvik sawmill is that it is absolutely necessary to continuously monitor the process performance to be able to see the effect of different technical changes. Above we have described how we can monitor the process performance by measurement of the measurement accuracy. This is of course a very important process parameter that has helped us to improve our process. However, there are also other parameters such as yield and stoppage frequency that need to be monitored in order to obtain a high efficient process performance. To conclude: Measure is to know in the daily improvement work. REFERENCES 1.

Grönlund, A. ; Flodin, J. ; Wamming, T. 2009. Adaptive control of green target sizes. In Proceedings of 19th International Wood Machining Seminar, Nanjing, China Oct 21 – 23, 2009.

2.

Grönlund A, (1992) Sågverksteknik del II, (in Swedish). X-725 1-1 000.1992.08 ISBN 91-7322-150-3

153


154

Process Monitoring Poster Presentations


Development of Instant Identification Method of CCA-Treated Wood using Laser-Induced Breakdown Spectroscopy Aono Yurie, Ando Keisuke, Hattori Nobuaki1 1

Tokyo University of Agriculture and Technology, Tokyo, JAPAN

ABSTRACT Laser-induced breakdown spectroscopy (LIBS) was applied to distinguish chromated copper arsenate (CCA)-treated wood from other treated wood. This technique has a potential to apply to in-situ analysis because it requires no sample preparation and can perform instant analysis. LIBS was applied to a surface of lumber with Nd:YAG laser emitting a pulse of 1064 nm in wavelength, 55 mJ/mm2 in fluence and 4 ns in pulse duration which is enough to create plasma. Fluorescence from plasma was collected by an ellipsoidal mirror, and then analyzed by spectrometer in the range of 200-300 nm. The results showed that the 228.9 nm As line and the 267.8 nm Cr line are useful to separate CCA-treated wood from non-CCA-treated one. Based on the elemental composition analysis by X-ray fluorescence, it was revealed that LIBS can identify CCA-treated wood accurately.

I#TRODUCTIO# CCA is soluble wood preservative which contains copper, chrome and arsenic. In Japan, three types of CCA were specified by Japanese Industrial Standard. The quality of CCA preservative is shown in Table 1. CCA impregnated lumber had been used from early 1960s. Demand for them increased rapidly Table 1. The quality of CCA preservative (JIS K 1570) since they have a high preservative Type 1 Type 2 Type 3 performance and low leaching rate. Chromium 59-69 33-38 45-51 From 1989 to 1995, 80% of treated compound wood were CCA [1]. Whose Content rate of Copper production amount is shown in effective compound 16-21 18-22 17-21 compound Fig.1. The total amount from 1965 (wt%) Arsenic to 2001 is estimated to about 8 15-20 42-48 30-38 compound million m3 [2]. The problem with CCA-treated wood is to release harmful gas which includes AsO2 [3] and ash which includes Chromium (VI) compound when they are incinerated [4]. From 1997, amount of the products decreased rapidly by voluntary action of wood preservative industry. Though the current production volume is very little, waste CCA-treated wood is being generated from demolition site. Estimated amount of disposal shown in Fig. 2 will reach its peak at 130 thousands m3 in 2015 and then decrease gradually. It continues for the next generation. Act for recycling of building materials established guidelines that CCA-impregnated wood must be separated and treated properly. However, the method which can identify them accurately at demolition site is still unestablished.

155


total for building material

140 Estimated amount of waste CCA treated wood (thousand m 3)

Amount of CCA -treated wood production (thousand m 3)

500 400 300 200 100

100 80 60 40 20 0

0 1965

120

1975

1985 Year

1965 1995 2025 2055 2085 2115 2145 Year

1995

Fig.1. Amount of CCA-treated wood production

Fig.2. Estimated disposal amount of CCA-treated wood

Therefore, laser-induced breakdown spectroscopy (LIBS) was applied to provide qualitative and quantitative data of the elemental composition of materials. In LIBS, an excited plasma is produced by a pulsed laser at a focal point. The plasma light is collected and analyzed by a spectrometer. The characteristics of LIBS are instant analysis, free sample preparation and simple operation. Several studies of applied LIBS for treated-wood analysis have been reported [5, 6, 7], but no one could develop a reliable identifying method or a portable equipment. In this study, we achieved to identify CCA-treated wood from non-CCA-treated one such as AAC, BAAC, ACQ, CUAZ even by low fluence pulse. Sampling method and accuracy of identification were also discussed.

EXPERIME#TS LIBS EQUIPME#T The LIBS analyzer used in this study is illustrated in Fig. 3. The laser source was a Nd:YAG laser (New Wave Research, Tempest 10) emitting a pulse whose wavelength, fluence and pulse duration are 1064 nm, 55 mj/mm2 and 4ns, respectively. A pulse was focused on surface of a sample by a plane-convex lens of f = 150 mm. The spot diameter was 0.7 mm. An ellipsoid mirror is used for collect plasma light in order to collect weak light efficiently. The collected light was led to a small spectrometer (StellarNet, EPP2000HR) whose resolution and analysis range were 0.15 nm and 290 – 300 nm, respectively.

SAMPLES Wood samples measured in this experiment are shown in Table 2. All samples were kept in a same room and air-dried before analysis.

156


Lens Laser

Ellipsoidal mirror Sample

Synchronous circuit Spectrometer Optical fiber

Fig. 3. LIBS analyzer

Preservative Chromium, Copper,

Ammonium Boron, Ammonium Copper, Ammonium Copper, (Boron) Unkown Non-treated

Table 2. Condition of samples Name Category Application Number CCA Unused wood 3 Waste wood Sill, Sleeper, Plinth, 11 Exterior wood AAC Waste wood Sill 3 BAAC Waste wood Sill 2 ACQ Unused wood 2 CUAZ Waste wood Unknown 2 Waste wood Sill 5 Unuesd wood 2

DATA ACQUISITIO# LIBS spectrum is usually affected by background light, pulse energy, distance between a lens and a sample. In addition, the heterogeneity of waste wood has a great effect on the spectrum. To minimize these effects, data processing is needed. Therefore, correction of baseline for all spectra was followed by normalization of all spectra by emission intensity of C 247.8 nm which was used for internal standard. Furthermore, to minimize effects of stain on sample surface and variation of preservative concentration in sample, averages of peak intensity at third laser shot for three different spot were used for identification.

RESULTS A#D DISCUSSIO# Distinct peaks of As were observed at 228.9 nm and that of Cr were at 267.89 nm. Fig. 4 show spectra of all the samples near As 228.9 nm and Cr 267.8 nm emission line, respectively. Strong peaks of As and Cr were observed in spectra from CCA-treated wood and not

157


observed from non-CCA-treated wood. Therefore, the thresholds can be set between the minimum emission intensity of As 228.9 nm and Cr 267.8 nm of CCA-treated wood spectra and the maximum of those of non-CCA-treated one. As shown in Fig.5, CCA-treated wood can be separated from others by these thresholds. The LIBS results were confirmed by X-ray fluorescence analysis to enhance reliability. In this study, the fluence of a pulsed laser was set to 55 mJ/mm2 which can be emitted even by a small laser. This suggests a potential of developing portable LIBS analyzer to identify CCA-treated wood.

3000 CCA ACQ BAAC ACQ CUAZ Unknown Non-treated

2000

150 1000

0

0

227

228 229 230 Wavelength (nm)

266

267 268 269 Wavelength (nm)

Fig. 4. LIBS spectra of all samples

3000 Cr 267.8 nm relative emission intensity (a.u.)

Relative emission intensity (a.u.)

300

CCA AAC BAAC ACQ CUAZ Unknown Non-treated

2000

1000

0 0

150

300

As 228.9 nm relative emission intensity (a.u.)

Fig. 5. Identification of CCA-treated wood

158


CO#CLUSIO# As LIBS could identify CCA-treated wood from other treated wood such as AAC, BAAC, ACQ, CUAZ by irradiating a pulsed YAG laser with the fluence of 55 mJ/mm2, this analyzing method is suggested to have a potential to use for instant identification of CCAtreated wood at a site. Further study is needed to investigate the effect of other analysis factor such as spot diameter, sample condition to develop a portable handheld LIBS analyzer.

REFERE#CES 1.

Iwasaki Katsumi (2008) BeikokunojuutakusyuuhengaikoubuzaisijoukaraCCAkeihozonsyorimokuzaigahaijosar etahaikei. Wood Preservation, 34(1): 2-12.

2.

Japan Wood Preservative association (2003) Kankyouniyasasiimokuzaihozonsyorigijutunokaihatsuhoukokusyo

3.

Cherles K. McMahon, Parhsall B. Bush, Edwin A. Woolspn (1986) How mush arsenic is released when CCA treated wood is burned? Forest Products journal, 36(11/12): 45-50Szymani, R. (1984) Using computers to design, operate and maintain saws, World Wood, 25 (8): 28-31.

4.

Japan Wood Preservative association (1985) CCAsyorihaizaianzenhaikisyorisuishinjigyouhoukokusyo – CCAhaizaianzensyoriiinkai-

5.

T. M. Moskal, D. W. Hahn (2002) On Line Sorting of Wood Treated with Chromated Copper Arsenate Using Laser-Induced Breakdown Spectroscopy. Applied Spectroscopy, 56(10): 1337-1344

6.

A. Uhl, K. Loebe, L. Kreuchwig (2001) Fast analysis of wood preservers using laser induced breakdown spectroscopy. Spectrochimica Acta Part B, 56(6): 795-806

7.

Takahashi Touru, Tomita Keiichi, Wakasugi Motoomi (2009) Development of Distinction Process of CCA Treated Wood from House Demolition using Laser Induced Breakdown Spectroscopy method. Hokkaidouritsukougyoushikenhoukoku, 308: 33-39.

159


Detecting Board Defects With a Time of Flight Camera Björngrim Niclas1, Hagman Olle1, Jonze Emil2 1

Luleå University of Technology 2Adopticum Center for Industrial Optics

ABSTRACT This study examines the suitability of using a Time of Flight (ToF) camera as a sensor for detecting saw defects such as saw mismatch, wedge, taper and snake. As demonstrator a warped board with saw mismatch was used. The results showed that the ToF camera could capture the geometries of the board. When looking at the saw mismatch the measurement accuracy is poor, the actual mismatch were 1,2 mm and the mismatch measured at two points with the ToF camera were 3,4 mm respectively 4,2 mm.

INTRODUCTION Raw material cost stands for about 70% of the total cost for sawmills. Utilizing the raw material to get as high yield as possible is of high importance [1]. Ideally a board should have parallel surfaces and a rectangular shaped cross section. However boards sawn with double arbor saws sometimes differ from this ideal shape resulting in mismatch between the upper and lower part of the sawn surface of the board [2]. A sensor that detects and monitor different shape defects of the sawn timber and also reconnects to the sawing process can help improve the performance of the process. The purpose of this report is to evaluate the suitability of a Time of Flight camera as such sensor. A board with warp and saw mismatch was used as demonstrator. Defects as snake, taper and wedge has not been studied, but if the camera can capture the character of the warp it’s reasonable to assume that it can capture the characteristics of the shape defects. A ToF camera generates a point cloud of the depicted object, where each point in the point cloud has X, Y and Z coordinates. With software as Matlab the points can be visualized and analyzed. Besides the obvious advantage of three-dimensional imaging compared to traditional cameras the ToF cameras are insensitive to ambient light conditions. ToF cameras are quick, and can take up to 100 frames per second, and they are robust since they don’t have any moving parts. All these properties could make it suitable as a fast non-contact method for on line measuring of board defects in sawmills. Saw mismatch, wedge, taper and snake are typical threedimensional problems, and the rapid development of 3-D techniques should make ToF a feasible solution both economically and performance-wise as a sensor for detecting these defects.

160


MATERIALS AND METHOD Material The ToF camera used in this study is a Fotonic B-70. Specifications of the camera are shown in table 1. The camera was mounted on a tripod for all the pictures taken. The camera needs to be coupled with a laptop to access the settings for the camera and in order to access the software viewfinder. Table 1. Specifications of the Fotonic B-70 Time of Flight camera [3]. Imaging parameters Max frame per second 50 Pixel array size 160x120 Pixel size 50x50 µm Aperture f 1,2 Distance accuracy (Z-direction, single pixel accuracy +/- 5mm at 0,1-1,5m +/- 10mm at 1,5-3m Distance uncertainty (frame to frame) +/- 5mm or 5% at 0,1-3m Illumination Illumination Laser 2x2W Wave length 808 nm Modulation frequency 44 MHz A time of flight camera shares some features with a traditional digital camera, but what is unique for the ToF camera is the use of modulated light and the pixels on the sensor have two receptors. The Fotonic B-70 ToF camera is equipped with two 808 nm laser diodes, which is modulated at 44 MHz. Each pixel on the sensor have two receptors, one receptor is in phase with the modulated light and one that is out of phase with the light. When taking a depth image the laser diodes oscillate several hundred of thousands of times per frame taken. Each light pulse has the light on for exactly the same time as the light is off, and the change between on and off is almost instant. The in phase receptor of the pixel is in exact synch with the light pulse and the out of phase receptor is out of sync with the light pulse. The light from the laser diode is modulated at 44 MHz. Each light pulse lasts for half a light pulse cycle, i.e. 88.000.000 light pulses per second. The speed of light is a constant, 299.792.458 m/s; in 1/88.000.000 s the light will travel 3,41m. If an object is located 1,705 meter from the camera the light pulse will travel 1,705m and then be reflected back to the camera for a total distance of 3,41m, and the time the pulse traveled is 1/88.000.000s, exactly in phase with the out of phase receptor of the pixels. Analogically will an object which are infinitely close to the camera reflect all the light to the in phase receptor [4]. One pine board with warp defects such as twist, bow, crook and saw mismatch and where used as demonstrator. The saw mismatch at the depicted area (see figure 1) where measured with a swage gauge, and maximum mismatch where 1,2 mm.

161


Method Both the saw mismatch and the warped board were depicted with 20 consecutive images, where the images were collected during 400 ms. When depicting the saw mismatch the camera was placed perpendicular to the board at a distance of approximately 0,16 meter from the object. For the warped board the distance was 2,3 meters. The ToF camera were mounted on a tripod and all depth images were taken from the same position, this allows to take the mean value of each pixel to increase the accuracy. The ToF camera generates two images in portable gray map (.pgm) format, one intensity image and one depth image. The depth images are processed with Matlab to create a viewable image. The mean value for each pixel in the 20 images was calculated for increased accuracy. The background of the pictures was subtracted for better visualization of the interesting areas.

RESULTS Saw mismatch

Figure 1. The picture shows the saw mismatch. Figure 1 show the mismatch studied. In figure 2 the coordinates for four points are shown. The coordinates are grouped two and two along the X-axis and perpendicular to the saw mismatch. The pair of coordinates is located at X=7 and at X≈-30. For the two points on X=7 the difference in depth (Z value) is 3,4 mm and for the points on X≈-30 the difference is 4,2 mm. The measured value of the saw mismatch is 1,2 mm.

162


Figure 2. Picture of the saw mismatch from the ToF camera, with space-coordinates. Warped board A warped board (figure 4) was depicted to examine if its geometry could be captured with a ToF camera. The point cloud was fitted to planes to better illustrate the twist of the board, see figure 3. To further visualize the shape of the board normal vectors for each plane were added to the matlab image, see figure 5.

Figure 3A &B. A shows the point cloud of the warped board. B show the point cloud fitted to 30 planes.

163


Figure 4. Picture of the warped board.

Figure 5. Normal vectors showing the twist of the board.

DISCUSSION AND CONCLUSIONS The pictures presented in this study are representations of 20 images that have been averaged in to one image. If a ToF camera where to be used on line in a sawmill one would have to rely on the depth data from one captured picture. The measurement accuracy would be worse but the characteristics of the saw mismatch should be captured. The accuracy of the depth picture of the saw mismatch might have benefited if plane adaptations had been done on each side of the saw mismatch, which would have reduced the distance uncertainty of the ToF camera. The shape of the warped board is well depicted even though the board is represented by only approximately 150x10 pixels. So with the low-resolution sensor on the ToF camera we still obtain a good representation of the shape of the board. The resolution of the ToF cameras available today is not good enough to give measurements with millimeter precision of defects such as saw mismatch, wedge, snake, etc. but they can capture the shape defects of the boards. The development of 3D sensor techniques is rapid and the accuracy will surely improve in the future. The information from the ToF camera used in this study despite its rather bad accuracy should be good enough to reconnect to the saw process in real time and alert when the sawing process is not running optimally.

164


REFERENCES [1] Grönlund, A., Flodin, J., Vikberg, T., Nyström, J., Lundgren N. (2009) Monitoring lumber size, shape and mismatch in double-arbour saws-Development and validation of scanning equipment. Proceedings of the 19th International Wood Machining Seminar. [2] Rasmussen, H., Kozak, R., Maness, T. (2004) An analysis of machine-caused lumber shape defects in British Columbia sawmills. Forest Products Journal Vol. 54, No. 6. [3] Fotonic. (2010) Fotonic B70 Leaflet. http://www.fotonic.com/assets/documents/fotonic_b70.pdf Retrieved 2010-04-26 [4] Canesta Inc. (2008) Canesta 101 Introduction to 3D vision in CMOS. http://www.canesta.com/assets/pdf/technicalpapers/Canesta101.pdf Retrieved 2010-04-23.

165


FuLOG – Radio Based Data Logger For Integration In Production Processes Ahmad, Kaleem 1; Grotekemper, Michael 2; Riegel, Adrian 2 and Witte, Stefan 1 1

2

InIT – Institute Industrial IT, Lemgo, GERMANY University of Applied Sciences Ostwestfalen-Lippe, Lemgo, GERMANY

ABSTRACT The radio-based data logger is used for online monitoring of temperatures during varied bonding tasks of the wood and furniture industry. The geometrical shape admits an easy integration in work pieces, for example in needed boreholes. According to this, the integration of an online monitoring system into continuous-flow machinery or to monitor series processes without timeconsuming preparations or costly appliances is possible. These are advantages over the up to date used methods, which are used only in special cases or in an indirectly way at particular points. Several scientific experiments proved a correlation of contact temperature and bonding strength for adhesive applications. The measurement of the temperature, which is reached during the first contact between the boundary layer and the work piece until the cooling, and a systematically interpretation allows a comprehensive quality control. Subsequently this will be the base of a quality control conception for bonding applications.

INTRODUCTION Hotmelts are employed in many category groups of the wood and furniture industry with varied bonding tasks. The motive to use hotmelts is the omitted drying time. The adhesive hardens within a few seconds of cooling. For this reason the hotmelts or by heat activated bounding systems are used for several continuous-flow machineries. Several research projects at the University of Applied Sciences Ostwestfalen-Lippe identified a directly effect of the adhesive temperature during the process to the quality of the bonded joint. Figure 1 shows the results of experiments with varied parameters to identify the most important process parameter to assure a qualitatively coating adhesion. The adhesion between the joint part and hotmelt requires a mechanical linkage of the hotmelt in the bonded part. The melt viscosity correlates directly with the melt temperature and the temperature of the joint part. In function of these temperatures the temperature of the boundary layer is reached during the first contact, which correlates with the bounding strength to an extremely high degree [1]. By heat activated bounding systems realise their characteristics like heat rack, wettability and initial strength only within a small range of temperature. This range of temperature must be reached at the first contact and must be hold during the forming unit. The adhesion strength will not be reached, if the adhesive is cooling down too fast after the first contact with the joint part. On the other side means an overvalued applied adhesive temperature a “swimming” of the coating, because the coating does not adhere within the forming unit.

166


Experimental Parameters: Maschining Centre Homag BAZ 41 Joker Adhesive EVA Jowat 280:30 Substrate MDF, edge shaped Testing Method: DIN EN 1464: Determination of peel resistance of adhesive bonds – Floating roller method

Figure 1: Effects of Varied Parameters during the Edge Banding Process The most important questions of adhesive applications are:  What is the temperature of the joint part and the adhesive?  What is the temperature at the first contact of the joint part with the adhesive?  What is the softening temperature in comparison to this?  What is the temperature during and after the forming unit? According to this the monitoring of the adhesive temperature is importantly and meaningful to decrease wastage and complaints, but this does not happen.

STATE OF TECHNOLOGY Up to now there are only measurements of adhesive temperature before the first contact with the joint part. These sensors are installed on the machinery and cannot detect the temperature of the boundary layer [2]. The monitoring of the boundary layer temperature occurs in praxis only in special cases or in an indirectly way at particular points and not the work piece itself (Solutions of Henkel AG and Jowat AG for process optimization for clients). The essential process steps like the heating and cooling of the adhesives are not recorded on a regular basis. For this reason it does not exists an all-embracing conception of quality control for a universal application of by heat activated adhesives. A similar approach is followed in relation to produce panel-shaped wooden materials. The solution herein consists of a miniaturised data logger for measurement of temperature and gas pressure inside the wood particle board. But in that case the sensor is executed as a lost component and his dimensions are too large for actual applications in furniture industries [3]. The University of Applied Science Ostwestfalen-Lippe has a solid experience in adhesive applications and the measurement of process parameters, especially the measurement of temperatures of laminating procedure (3D-coating) [4][5], edge bending [6][7] and profile wrapping [8]. These analytical and empirical studies are the base of coming studies to establish a concept for quality control.

167


AMBITION OF “FuLOG” Conclusions like they are used today are shown in Figure 2. The volition to monitor the temperature online during the process means a special, time consuming preparation of a work piece (connecting sensors, wiring and assembly of measurement technique). In consequence of this, the monitoring is rarely used.

Workpiece Rest Press Table

Figure 2: Measuring Technique during Membrane Pressing The ambition of the project “FuLOG” is the development of a miniaturised, energy self-sufficient data logger for simply integration in work pieces. The communication for parameterisation, data exchange and triggering occurs radio-based during the process. The data exchange can also occur after the process depending on storage of data during the process. The geometrical shape admits an easy integration in work pieces, for example inside of needed boreholes. Figure 3 shows a schematic representation: This concept affords important advantages:

Fu LO G

 The recording of measuring data inside the logger (without transmitting during the measurement) admits high sample rates, which are not based on the communication characteristic of the transmission path. So the monitoring of high dynamically processes are even possible.

Fu LO G

 The time and material basis for a preparation of measuring is decreased to a high degree. There is also no special tool machine for placing the “FuLOG” required. This admits even a frequent monitoring to improve quality and processes.

IR Heating Unit Temp. Of First Pressure Cooling Curve

Temperature

 The use of thermocouples means no thermal -- Temperature capacity of the temperature measurement measured by instrument, which would falsify the results. FuLOG They are low-cost parts and remain within the Time boundary layer. Figure 3: Possible Integration of FuLOG during the Profile Wrapping Process

168


The Idea of the miniaturised, wireless data logger is based on the experiments to measure the temperature during adhesive applications within the wood and furniture industry. But the concept of “FuLOG” is based on a basic module enclosing the energy supply and the data logger. According to the measuring task diverse sensors get linked on this module. For this reason the advantage of “FuLOG” is the omitted wiring and space saving placement next to the place of measurement.

TECHNICAL DETAILS OF “FULOG” “FuLOG” employs an MSP430, a 16-bit microcontroller equipped with a 12-bit analog-to-digital convertor (ADC) for temperature readings. Radio interface consists of a CC2500 transceiver to support wireless data transfer to a central data logging server. All components of “FuLOG” are arranged in a two layer architecture, where battery, antenna and transceiver are placed at one layer, while all other components are placed at the other layer as shown in Figure 4. The layered architecture minimizes the size of device at one hand and offers convenient adaptation of important parameters at the other hand. Hence, only layer A, which consists of energy and transmission related components, needs to be changed in order to change the operating frequency or the battery life of “FuLOG”. Meanwhile “FuLOG” can be tuned to operate in the 2.4 GHz or 915 MHz frequency bands without significant changes in the software. The default operating band is 2.4 GHz band. Main features of “FuLOG” are listed in Table 1. Moreover the “FuLOG” device contains a 512 kB on board memory, which enables it to store the measurements for several minutes, depending on the configured sample rate. Hence engineers can either examine the logging data at once at the end of an experiment by manually connecting the device with logging server using a USB interface or can transfer this data to the logging server using the radio interface, even while the experiment is still in progress. The logging server can control measurements and logging operations of all “FuLOG” devices in the network either individually or collectively depending on the choice of the user. Maximum three temperature sensors can be connected to a single “FuLOG” device, hence one device can take readings of three different spatial points simultaneously. Crimping based SMT wire-to-board connectors are used to connect temperature sensors with “FuLOG”. It makes it easier to connect the delicate and extremely thin sensor wires and eradicates the need of screwing. Furthermore a rechargeable battery is used, which reduces the cost of operation. Layer B

Maximum three temperature sensors

Layer A

LIR2032 Battery 35 mAh @ 3.6 V 26 mm

CC2500/ CC1100 Radio 2.4GHz/915MHz Antenna

Figure 4: Two Layer Architecture of FuLOG

169


Table 1: Specification of FuLOG Operating frequency Power consumption Battery life Maximum no. of sensors Size Temperature range Memory

2.4 GHz, 915 MHz 20 mA 105 min 3 26 mm x 12 mm -40°C ... 60°C 512 kB (changeable)

Even in application fields wherever local measurements are required and have to accomplish rapidly thermocouples are suggested. Because of their low costs, simple manufacturing and handling, thermocouples are favoured in application fields, where mechanical effects destroy the sensors. In case of the adhesive applications within the wood and furniture industry the sensor (thermocouple) is bonded between the joint part and the coating, consequently the sensor is waste.

CONCEPTION OF A QUALITY CONTROL SYSTEM Based on the analytical and empirical studies of processes using hotmelts or by heat activated bounding systems and the use of “FuLOG” permit a conception of a quality control system for the furniture industries. The following Figure 5 shows a general process illustration using bounding systems:

Figure 5: Conception of a Quality Control Management

170


It is a fact that the temperature of the bounding layer correlates directly with the bounding strength to a very high degree (compare Figure 1) [1]. Hence the need, firstly, to define quality parameters, which are achieved at defined temperatures. These quality parameters have to be defined with the aid of destructive testing methods. Furthermore, the relevant data of machine supervisors and the data of materials at the beginning of the process have to be saved in a database. Based on this information and the measurement of process parameters, especially the monitoring of the temperature of the boundary layer by “FuLOG”, permit the machine supervisor to gain direct influence on the process. Nowadays the machine supervisor adjusts the process to the best of his knowledge. After the testing of the adjusted parameters he studied the result and change one or two parameters and tests again. This procedure is repeated until the result complies with the quality requirements and the production process itself starts. This time-consuming set up has to be avoided and even more, because it is not guaranteed that the same quality is reached with the same parameters during a following production. The aim of a concept of quality control system should not be based on a self-adapting automatized control loop. The concept should generate additional information for the machine supervisor for regulation but also a logging of the chosen parameters for following productions. This would cause a reliable production based on the proposal of the system from the outset.

REFERENCES 1.

Hoffmeister, H-W.; Horstmann, S.; Riegel, A.; Strauß, H. (2007) Abschlussbericht AiF 14525 N.

2.

Horstmann, S.; Riegel, A. (2008) Ummantelungsverfahren besser verstehen. Prozesssicherheit beim Fügen. Adhäsion 52 (3): 28 – 34

3.

Hasener, J. (2010) Mittendrin…statt nur dabei; Temperatur und Gasdruck in der Heißpresse drahtlos messen. MDF-Magazin 16: 42 – 45

4.

Stüttgen, B. (2006) Beitrag zur Ermittlung des Wärmestandes an folienkaschierten Möbelfronten, Diploma thesis at the HS-OWL University of Applied Science, Lemgo, Germany

5.

Fuchs, I.; Wenk, S. (2010) Einflussgrößen auf die Prozesssicherheit bei der 3DBeschichtung von Möbelfronten. Holztechnologie 51 (3): 34 - 38

6.

Wittenstein, K. (2007) Optimierung des Kantenanleimens auf Bearbeitungszentren. Final report TRAFO

7.

Riegel, A. (2006) Temperaturverlauf bei der Schmalflächenbeschichtung. Holz als Rohund Werkstoff 64 (5): 431-432

8.

Horstmann, S.; Riegel, A. (2008) Temperaturverlauf des Schmelzklebstoffes beim Ummanteln. Gegenüberstellung von näherungsweisen Berechnung mit OnlineMessungen. Holztechnologie 49 (3): 26 – 30

171


172

6. Sanding - Surface Evaluation Oral Presentations


Modeling wood surface geometry after wood machining Sandak Jakub Trees and Timber Institute IVALSA/CNR, San Michele All’Adige, ITALY,

[email protected]

ABSTRACT Wood surface smoothness generated during wood cutting is one of the most important indicators quantifying process quality. It affects both aesthetical properties and future performance of wood products. The surface is usually generated by the cutting tool. However, a composite structure of wood, as a biological material, makes the surface geometry very complex. The goal of this work was to attempt development of numerical modeling method for simulation of the surface formation while cutting wood. Several peculiarities of the wood cutting process have been considered and implemented in to algorithms, such as wood anatomical structure, cutting geometry, cutting angles and tool wear.

INTRODUCTION Every material object has its surface composed of miniature peaks and valleys. The size and spatial distribution of these peaks/valleys influence the specific surface properties (both aesthetical and future performance). The problem of wood surface smoothness is especially complex and therefore it has been already studied. An attempt to summarize some of the problems related was presented for instance within the previous work of author [1]. The work presented here is its continuation and extension.

dynamic

machine,

properties

tool

material

manufacturing

properties

process

workpiece

SURFACE

generated

METROLOGY

Figure 1 Factors influencing generation and evaluation of surface smoothness Usual methods of processing wood base on different variations of cutting. Various anatomical elements are cut by the tool during machining, and a very complex surface is created as a result. Moreover, the anisotropy of wood, wood density, moisture, along with the kinematics of the cutting process, machine conditions and other variables enhance the complexity of the wood

173


surface. Figure 1 presents flowchart of interactions between various factors affecting generation of the wood surface relief. This schema is inspired by the work by Whitehouse [2] developed for isotropic metal-type materials. In contrast, the effect of wood microstructure has a great effect on generated workpiece surface. For that reason, a thick arrow has been added to the original Whitehouse’s graph. The surface geometry of wood can be considered as a superposition of various sub-geometries related to diverse factors [1, 3, 4, 5]. Some of these result from the manufacturing process, other are consequences of the dynamic relation with the tool-workpiece. Many wood surface irregularities are an effect of material properties and microstructure. A number of the most important issues affecting the wooden surface form are [1]: • Wood anatomy • Grain figure • Wood density/wood porosity • Moisture • Kinematics of the cutting process • Machine conditions o design of the machine o machine vibrations o tool wear o tool maintenance (quality of sharpening, joining of the cutting knifes) o other factors; such as strength of the positioning of the workpiece and stiffness of the tool holder • Other factors o temperature o air humidity o surface finishing (lacquer, paint) o chemical decomposition of the surface (oxidation and degradation) o biological rotting (brown, white or grey decay) o damage of the surface as a result of insect activity It is possible to analyze an affect of each of the above factors separately. Several researches have been dedicated to estimate effects of wood species [6, 7] machining parameters [8, 9] among others. On the other hand combining all the factors together would provide much better understanding of the surface formation and it properties. The logical consequence is therefore development of numerical models. It must be mentioned that some attempts for modeling of the machined surface have been reported [10, 11, 12, 13], but all these were related to metal.

OBJECTIVES The goal of this work was to develop a novel numerical modeling tool dedicated to simulation of the surface formation while cutting wood. The model bases on the virtual processing by means of 3D solid editing and other numerical methods. Several peculiarities of the wood cutting process should be considered and implemented in to algorithms, such as wood anatomical structure, cutting geometry, cutting angles and tool wear.

174


GENERAL CONCEPT The core idea of the model bases on the “virtual cutting” approach as summarized in Figure 2. Three main modules are critical for the model performance: 1. generation of virtual workpiece (including peculiarities of wood, such as anatomy, variations of the cells in radial an tangential direction, defects, etc) 2. creation of virtual tool path (including all disturbances such as unbalance, vibrations, deflections, etc.) 3. analysis of the surface generated (including estimation of roughness parameters, effects of filtering, 3D pattern, etc.) Generate “virtual” material 2D map

Generate “virtual” cutting edge path

Create solid

Include vibrations and distortions

Rotate according to cutting direction

Create solid

Subtract both solids

Analyze geometries Figure 2 Virtual cutting algorithm

METHODS Generation of wood cellular structure In order to model an effect of wood anatomy, a dedicated software module was developed mimicking wood anatomical structure. Wood was considered here as matrix of elongated empty tubes, with varying dimensions in radial and tangential directions. Several parameters have been identified and implemented in to software. Figure 3 presents a schema of the simplified softwood cross section.

175

D


E

F C

B A

Figure 3 Definition of parameters characterizing anatomical structure of softwood; A – yearly ring width, B – early wood tracheid width, C – late wood tracheid width, D – tracheid height, E – early wood cell wall thickness, F – late wood cell wall thickness Two different approaches for cross section modeling were defined and implemented in to software modules: a. generation of the cross section on the basis of statistical data b. generation of the cross section on the base of microscopic images of wood An example of the softwood cross section generated by the software on the basis of the statistical parameters (as defined in Figure 3) is presented in Figure 4. In that case the microstructure was considered as a matrix of rectangular cells changing its dimensions in radial and tangential directions. The random variation of each cell was also introduced for all parameters (from A to F). The code was generated in LabView. As a result of computation coordinates of lumens were saved in a text file.

Figure 4 Print screen of the software module generating 2D structure maps (cross section) of softwoods In the following step the text file was opened by Visual Basic Script of AutoCad. The coordinates were used for creation of quadrangles, than regions and finally extruded in to 3D solids. As a

176


result a three dimensional virtual workpiece was constructed. The solid was rotated later, depending on the desired cutting direction. Generation of the cutting edge path The cutting edge was modeled as a line, however it is possible to introduce additional variable to the model related the cutting edge shape deviations (profile, tool wear, catastrophic edge recession). The movement direction of the edge was computed depending on the movement vector defined in various modes, as shown in Figure 5. The position of the cutting edge was updated every fraction of time dt, depending on the desired resolution.

a)

b) v(t)

c)

v2(t)

v(t) v1(t)

Figure 5 Representation of the cutting edge and movement vectors; linear cutting without any disturbances (a), linear cutting with some disturbances common to the whole edge (b), linear cutting with different disturbances on both sides of the cutting edge (c) Customized cutting edge paths were implemented for different types of wood machining process. An example of the rotary cutting (such as planning) is presented in Figure 5. The left side of the cutting head was not vibrating at all; however the vibration of right side were defined as vertical mode of known vibration frequency (fv), phase shift (φv) and vibration amplitude (Av). As expected the cutting edge paths differed in both sides. The other parameters to be provided before computation include cutting rotational speed (n), feed speed (vf) and tool diameter (D). The length of the path was limited by the time of machining (T), and resolution by delta time (dt). Some mathematical formulas (in relation to rotary cutting such as planning) utilized for algorithm inside of LabView as presented in Figure 6were as follows: •

horizontal position of the rotation centre;

•

vertical position of the rotation centre;

•

horizontal position of the cutting edge;

•

vertical position if the cutting edge;

x0 (t ) = v f ⋅ t

y0 (t ) = sin( f v ⋅ 2π ⋅ t + ϕv ) ⋅ Av D xe (t ) = x0 (t ) + sin(n ⋅ 2π ⋅ t ) ⋅ 2 D ye (t ) = y0 (t ) + cos(n ⋅ 2π ⋅ t ) ⋅ 2

Results of the computation were saved as txt file and forwarded to the Visual Basic script of AutoCad, in analogy to the virtual workpiece method described before. The region creation and extrusion was performed in order to generate a 3D solid representing volume (equivalence of chips) removed by the cutting edge. The virtual surface was generated after extraction of the 3D solid of the cutting edge from the 3D solid representing workpiece.

177


Figure 6 Print screen of the software module generating cutting edge path (including vibrations - Right and without vibrations - Left)

EXAMPLE OF UP-TO-DATA RESULTS Several sub-tasks of the project have been already accomplished, however it is still under development, as mentioned before. An example of the modeling software potential is represented on Figure 7. A piece of softwood of defined anatomical structure was virtually processed in rotary cutting (routing) by perfectly sharp tool. The resulting surface geometry and magnification of the part is plotted in Figure 7. A very detailed reproduction of the structure was obtained.

Figure 7 Example of the virtually machined surface generated with the developed software (softwood, routing, tool diameter 50mm, feed per tooth 2,5mm, feed direction 1° to the fiber direction)

178


FUTURE WORK It is clearly missing for the moment a module quantifying numerically modeled surfaces (in terms of surface roughness parameters). Such software is already under development and will be included to the package in a close future. Other works planned include: • Module digitizing microscopic cross-section images of real wood samples and generation of the virtual workpieces on the base of such images (especially hardwoods) • Implement different cutting techniques in to virtual machining (such as sawing, sanding, peeling, etc.) • Model accuracy of selected sensors utilized for wood surface roughness measurements (such as styluses of different geometries, optical sensors, etc) • Extensively explore virtual surfaces in order to understand “how much roughness on wood surface is from biology and how much from engineering”

ACKNOWLEDGMENTS Part of this work has been conducted within framework of SWORFISH (Superb Wood Surface Finishing) project co-financed by Provincia Autonoma di Trento.

LITERATURE 1.

Sandak J. and Negri M. 2005 Wood Surface Roughness – What Is It?, Proceedings of the 17th International Wood Machining Seminar, Rosenheim, Vol.1 pp.242-250 2. Whitehouse DJ (1997) Surface metrology. Meas. Sci. Technol. 8:955-972 3. Riegel A (1993) Quality measurement in surface technologies. International Conference on Woodworking Technologies, Ligna, Hannover pp 23:1-9 4. Ohtani T, Tanaka C, Usuki H (2004) Comparison of the heterogeneity of asperities in wood and aluminum sanding surfaces. Precision Engineering 28(1):58-64 5. Thomas TR (1999) Rough surfaces. Imperial College Press, London. 278p. 6. Magoss E., Sitkei G. and Lang M. 2005 New Approaches in the Wood Surface Roughness Evaluation Proceedings of the 17th International Wood Machining Seminar, Rosenheim, Vol.1 pp.251-257 7. Fujiwara Y., Fujii Y., Okumura S., 2003 Effect of Removal of Deep Valleys on the Evaluation of Machined Surfaces of Wood. Forest Products Journal pp. 58-62 8. Jackson M., Hynek P., Parkin R 2007 On planing machine engineering characteristics and machined timber surface quality Proc. IMechE Vol. 221 Part E: J. Process Mechanical Engineering pp.17-32 9. Orlowski K. 2010 The fundamentals of narrow-kerf sawing: the mechanics and quality of cutting. Publishing house of the Technical Univeristy in Zvolen, p.123 10. Adamczak S., Miko E., Cus F. 2009 A Model of Surface Roughness Constitution in the Metal Cutting Process Applying Tools with Defined Stereometry. Journal of Mechanical Engineering 55(1):45-54 11. SchmitzaT., Coueyb J., Marshb E., Mauntlera N., Hughesa D. 2007 Runout effects in milling: Surface finish, surface location error and stability. International Journal of Machine Tools & Manufacture 47:841–851

179


12. Chang c., Lu H. 2006 Study on the prediction model of surface roughness for side milling operations. Int J Adv Manuf Technol 29:867–878 13. Rojas H., Corral I., Calvet J. 2007 Numerical simulation model for the determination of surface roughness in side milling as a function of feed and of the tool edges radii. Archives of Materials Science, Vol. 28, No. 1-4, pp. 51-55

180


Sanding with coated abrasives – Relations between temperatures, power and surface quality Hoffmeister, Hans-Werner, Luig, Martin and Fricke, Alexander Institute of Machine Tools and Production Technology (IWF), Technische Universität Braunschweig, GERMANY

ABSTRACT As sanding with coated abrasives is often the last operation in wood machining it is important for the final surface quality of the workpiece. The surface quality and the wear of the tool depend not only on the process and workpiece parameters but also on the specifications of the coated abrasives. However, there is no model which describes the relations between these properties. This results in a selection of sanding tools and process parameters based on empirical values which often lead to suboptimal parameter combinations and a lower productivity. Because of that experiments are made by the Institute of Machine Tools and Production Technology to clarify the relations and to build up a data basis. The topography of the coated abrasives was measured by a laser sensor and evaluated to show the grain density and the grain size distribution. Sanding tests with a belt sander were made to measure the tangential force, the material removal rate and the temperature distribution in the workpiece with different parameter combinations to calculate the process power and its components (cutting power to remove material from the workpiece and thermal power, which warms up the workpiece). In addition to that sanding tests were made to analyze the final workpiece topography using a confocal microscope.

INTRODUCTION The most important operation for surface finishing of solid wood products in the wood-working industry is sanding with coated abrasive belts. An overview of the different kinematics is given in DIN 8589 [1]. Belt sanding is a process with geometrical undefined cutting edges. The surface quality of the workpieces depends on the grain size, the grain density and the distribution of the grains [2]. There are two different effects which lead to a decreasing material removal rate over the sanding time depending on the workpiece material. The predominant effect if sanding solid wood or wood products with coated abrasives is, that the chip space is loaded up with chips. In this non-continuous process the loaded up structure gets instable which leads in a chip removal using the sanding process kinematics [3]. An abrasive wear effect to the grains of the tool can be noticed, if hardwood with silicates is sanded [4]. With measurements of the topography of the coated abrasive in discrete time-steps during a sanding test the tool life behavior can be shown. To measure the topography Schnettker [5] developed a system with a laser sensor and a cross table to move the sample. The results of the evaluation of the measured topography are the grain density and the three-dimensional material-curve (AbbottCurve). Although a lot of research work has been done in the past there are no results which can describe the relations between the given parameters (process parameters, workpiece parameters, specification of the coated abrasives), the resulting surface quality and the tool life behavior. Because of

181


that, sanding tests with different parameter combinations were made to clarify these relations. The results of these tests are shown in this article.

EXPERIMENTAL SETUP, WORKPIECES AND TOOLS For the sanding tests eleven different coated abrasives and two kinds of wood were used. The workpiece materials were beech tree and pine tree. Out of that workpieces were cut to realize a 140 x 70 mm contact zone between the workpieces and the coated abrasive. The coated abrasives had the grain size classes P60, P150, P220 and P320 with a closed (siawood 1920), half open (siawood 1919) and open (siawood 1939, excluding P320) grain density. The coated abrasives consist of a paper basis with a resinoid bond and aluminium oxide grains (P320 half open – silicon carbide). The setup of the sanding tests is shown in Figure 1.

Figure 1: experimental setup for sanding tests The left side in Figure 1 shows the test stand which is mainly based on a belt sander, a mass balancer and a linear guided workpiece. The transversal degree of freedom of the clamped workpiece is in vertical direction so that the workpiece clamping device weight applies a normal force onto the workpiece. This force can be changed by varying the mass balancer´s weight. On the right hand side of the picture the schematic setup to measure the temperature in the workpiece is shown. For the measurement of the temperature distribution in the workpiece six thermocouple elements were placed in small holes in the workpiece. There were two rows of holes with three different depths at the front and the end of the sanding zone (see Figure 1, right side). The tangential force between the workpiece and the coated abrasive was measured by a piezo-electric force transducer which is placed between the workpiece clamping device and the linear guidance. The inductive displacement transducer measured the reduction of the thickness of the workpiece. With this measured data it was possible to calculate the material removal rate. For the measurement of the surface quality sanding tests were made without thermocouple elements and an inductive displacement transducer. Moreover, the sanding process was interrupted in discrete timesteps so that the surface roughness could be measured using a confocal microscope. The sanding tests were done with three different cutting speeds and normal forces. The values of the varied cutting speed were 150, 225 and 300 m/min. The normal force had the values 43, 61

182


and 77 N. The workpieces were sanded using all coated abrasives mentioned above. In addition to that the wood fibre direction of the material was chosen against to and along to the direction of the cutting speed by sawing the workpieces. This was done to show the influence of the wood fibre direction but in an industrial sanding process it is not possible to choose the wood fibre direction in the most cases. During the measurement of the material removal rate and the workpiece temperature the moisture of the workpieces was also varied in a range from 10 % up to 20 %. However, there was no visible effect depending on this parameter in the varied range.

GRAIN DENSITY AND MATERIAL REMOVAL RATE The topography of the coated abrasives was measured with the measurement system which is shown in Figure 2.

Figure 2: Measurement system for coated abrasives The sample of the coated abrasive has to be fixed on the cross table. The laser sensor measures pointwise the distance between the sensor and the sample. The cross table moves the sample relative to the laser sensor so that parallel lines of measurement points of the sample can be measured. From these measurement data the topography of the coated abrasive is calculated. The topography is the basis to calculate the grain density and the three-dimensional material-curve. The grain density of the coated abrasives is shown in Table 1. The grain density increases with higher P-numbers and decreases from the closed to the opened grain structure. The exception at P320 is caused by the different grain materials of the coated abrasives. Table 1: Grain density of the coated abrasives P60 P60 P60 P150 P150 P150 P220 P220 P220 P320 P320 c h o c h o c h o c h grain density 4,0 2,8 2,4 [1/mm²] c = closed , h = half open, o = open

17,8

15,9

183

15,4

36,3

32,9

29,1

73,6

74,3


The material removal rate is defined as the speed of the thickness reduction of the workpiece and is calculated from the measured data of the inductive displacement sensor. The material removal rate is decreasing in the direction of rising grain numbers and it is increasing inside a P-number category from grain density closed to open. An exception is grain number P320 where the closed coated abrasive has a higher material removal rate because of the grain density and the grain material. Workpieces of beech tree have a significant lower material removal rate as workpieces of pine tree because beech tree is harder and has a higher strength than pine tree. Furthermore, sanding along the wood fibre leads to a lower rate in comparison to sanding orthogonal to the wood fibre. A higher normal force at workpieces of pine tree increases the material removal rate. However, the effect is particular nonlinear for beech tree especially using the grain number P60 and P150. Because of the hardness of the workpieces the engagement of the grains into the workpiece at the lowest normal force is very small. This leads to a small chip cross-sectional area. Because of this and the negative rake angle in combination with the high strength of the material the grains are particularly pushed out of the workpiece through the kinematic of the sanding process. This effect gets smaller with higher normal forces. This leads into an over-proportional increase of the material removal rate between the lowest and the middle value of the normal force. The material removal rate is proportional increasing with a higher cutting speed at both materials. In the most cases the material removal rate was independent from the sanding time. An exception could be observed if sanding workpieces of beech tree with the highest normal force and cutting speed. At these experiments the coated abrasives shows a loading of the chiproom.

TEMPERATURE DISTRIBUTION AND POWER COMPONENTS The main components of the process power Pp which is needed for the sanding process are the cutting power Pc to remove chips from the workpiece and the thermal power Pt. The relations between the power components and measured data are shown in equation 1 - 4. Pp  Pc  Pth (1)

PP  Ft  v c

(2) (3)

Pth  Pth , Wst  Pth ,Wz

T  m Wst  c Wst   Wst  (TWst  TU )  A Wst (4) t The process power is calculated with the measured tangential force Ft in the contact zone between workpiece and tool multiplied with the cutting speed vc (equation 2). The thermal power can be calculated using the temperature distribution change in the workpiece and in the tool (equation 3). The cutting power is the difference between process power and thermal power. The thermal power which heats up the tool (Pth,Wz) is neglected because it was not possible to measure the temperature distribution of the tool during the sanding tests. Furthermore, the mass and the heat capacity of the tool are low in comparison to the workpiece. The thermal power which heats up the workpiece (Pth,Wst) is calculated with the measured temperature difference ΔT in the workpiece multiplied with the mass mWst and the specific heat capacity cWst of the workpiece divided through time difference Δt between the temperature measurements (equation 4). The second term in equation 4 describes the power which is getting out of the workpiece caused by the heat transfer between the workpiece with temperature TWst and the environment with temperature TU over the face AWst of the workpiece. A typical temperature development within a workpiece during sanding beech tree with coated abrasives with high P-numbers is shown in Figure 3. Because of the limited heat conduction of the workpiece and the distance between contact zone and thermocouple element there is no variation in the temperature in the beginning of the measurement (Phase 1). After this phase there are two areas with a progressive and a linear increasing Pth , Wst 

184


temperature. In the last phase the temperature increase becomes continuous smaller because of the heat transfer to the air around the workpiece. When sanding pine tree or beech tree with low grain numbers this effect can´t be observed because of the higher material removal rates. During the sanding tests the distance between thermocouple element and contact zone gets significant smaller. In this case, there is no degressive increasing in the temperature development at these sanding tests.

Figure 3: Temperature development in the workpiece

Figure 4: Influences to the process power and the chipping power

The temperature of the workpiece is higher when sanding beech tree in comparison to pine tree. It is also higher along the wood fibre as orthogonal to the wood fibre. Furthermore, a higher moisture leads to a slightly increase of the temperatures because the heat conduction increases with a higher moisture. A higher normal force or cutting speed also leads to a higher temperature in the workpiece. The lower temperatures at sanding against the wood fibre are caused through the higher material removal rate which needs a higher cutting power at the same process power. In

185


addition to that a part of the thermal power is getting removed of the contact zone by the chips before it can get deeper into the workpiece. The process power for the whole measurement is calculated with equation 2. Because of the invariant temperature at the beginning of the measurement the thermal power and the cutting power are calculated for phase 3 and 4 of Figure 3 first with equations 4 and 1. Afterwards the cutting power and the thermal power at the beginning of the measurement were calculated with the results of phase 3 and 4, assuming that the cutting power is proportional to the material removal rate. The influence of the parameters to the process power and the cutting power is shown in Figure 4. The process power is proportional increasing with a higher cutting speed or normal force. The process power is proportional to these two parameters because it is calculated out of the product from the cutting speed and the tangential force. The cutting power is also proportional increasing with a higher cutting speed or normal force. The process power is higher when sanding beech tree and against the wood fibre in comparison to pine tree and sanding along the wood fibre. The process power is higher against the wood fibre because of the higher material removal rate. The power per removed volume is against the wood fibre lower than along the wood fibre. There is no significant difference between the cutting power at beech tree and pine tree, although beech tree is much harder than pine tree. However, if one brings the chipping power in relation to the material removal rate the chipping power per removed volume is higher for beech tree than for pine tree. When sanding orthogonal to the wood fibre the cutting power is higher than along the wood fibre because of the higher material removal rate. The cutting power per removed volume is lower against the wood fibre. This shows that it is easier to remove material from the workpiece against of the wood fibre. The process power and the cutting power decrease with a smaller grain size. The grain density shows no effect to the power components in the tested range. The moisture has no influence to the process power or the cutting power in the tested range with a moisture content between 10 % and 20 %, too.

SURFACE QUALITY To evaluate the surface quality of the workpieces the topography was measured with a confocal microscope after defined times during sanding. With this method it was possible to analyse the development of the surface quality over the sanding time to evaluate the influence of the wear of the tool. The measured lines were orthogonal to the direction of cutting with a length of 48 mm. The measured topographies was filtered with a high-pass filter (cut-off wavelength: 8 mm) and a low-pass filter (cut-off wavelength: 25 µm) to get the roughness profile from the measured profile . The roughness profile was evaluated to determine the roughness-parameters Ra, Rz, Rpk and Rvk which are indicators for the surface quality. The workpieces of pine tree had a high initial roughness with a high variation because these workpieces were premanufactured by sawing. The workpieces of beech tree had a lower initial roughness without a high variation because they were sanded with a rough coated abrasive. The roughness parameter Ra is rapidly falling during the sanding tests to a constant value which keeps constant till the end of the sanding tests as shown in Figure 5. This effect is slower at coated abrasives with small grains (P320) when sanding pine tree because of the high roughness of the workpieces at the beginning of the sanding tests. In addition to that the material removal rate is lower, which make the initial topography longer to be seen during process. There could be shown no effect to the tool wear during the sanding processes, yet. The roughness parameters have lower values when sanding with coated abrasive with a smaller grain size. An influence of the grain density within a grain size group could not be recognized during the sanding tests because the differences in the grain density weren’t very high. The roughness parameters Rpk and Rvk show

186


the same behaviour than Ra. The roughness parameter Rz is also falling over the sanding time to a constant value, but a longer sanding time is need to reach the constant value. The reason for this is single fibres which are face out of the material as a result of the sanding process.

Figure 5: Influence of the sanding time to Ra

Figure 6: Influence of the wood and the fibre direction to Ra

The roughness of the workpieces of pine tree is higher than at workpieces of beech tree if sanding with coated abrasives with low P-Numbers (here P 60). Furthermore, when sanding with such coated abrasives along the wood fibre the roughness is higher than against the wood fibre. An example for these influences is shown in Figure 6. The roughness of sanded pine tree is higher, because of the wooden structure and hardness. There are more fibres face out of the material and the grains of the coated abrasives get deeper into the workpieces than at workpieces of beech tree. The slightly lower roughness when sanding against the wood fibre is caused by the direction of the measurement. This is orthogonal to the cutting direction and therefore along the wood fibre. Consequently, there is a partial compensation of the roughness of the sanded structure through the roughness of whole removed fibres. The tested grain sizes P 150, P 220 and P 320 do not show an influence of the wood fibre direction, because there less whole removed fibres. With

187


smaller grain sizes the difference in the roughness between beech tree and pine tree becomes a constant value because the influence of the wood structure to the roughness increases if sanding with coated abrasives with smaller grain sizes and beech tree and pine tree have different wood structures. An influence of the cutting speed to the surface quality could not recognize during the sanding tests. The influence of the normal force is not analysed yet, but it is assumed that the surface quality depends on the normal force. Because of the higher material removal at higher normal forces it also is the assumed that the grains get deeper into the workpieces and therefore the roughness is higher.

SUMMARY AND OUTLOOK There are many parameters which influence the surface quality, the process power and the material removal rate when sanding wooden parts with coated abrasives. In this article the influence of the process parameters, the workpiece parameters and the specification of the coated abrasives to the material removal rate, the process power and the surface quality is shown. The results of the sanding tests show a high influence of the process parameters to the material removal rate and the process power. The material removal rate and the surface quality depend on the specification of the coated abrasives and the workpiece parameters. Further sanding tests are planned to show the influence of other workpiece materials like medium density fibreboard and of structural defects in solid wood (e.g. knots) to the material removal and the process power. Another research interest will be the evaluation of the wear of the coated abrasives from measurements with the introduced laser sensor and REM-pictures.

ACKNOWLEDGEMENT This work is supported by the German Research Foundation (DFG) under the numbers HO 1878/34-1 and OS 166/7-1. It is a project in cooperation with the Institute of Dynamics and Vibrations (IDS) of the TU Braunschweig which use the results of the sanding tests to build up a model based on cellular automats.

REFERENCES 1.

DIN 8589-12 (2003) Fertigungsverfahren Spanen – Teil 12; Einordnung, Unterteilung, Begriffe, Beuth-Verlag, Berlin

2.

Honig, H. (1999) Aktuelle Entwicklungen in der Holz- und Lackschlifftechnik, IndustrieLackierbetrieb, 67 (2): 80-85

3.

Westkämper, E,,Riegel, A. (1995) Standzeitverhalten von Schleifwerkzeugen beim Schleifen von Fichte, Holz als Roh- und Werkstoff, 53: 281-287

4.

Koch, P. (1980) Utilization of Hardwoods Growing on Southern Pine Sites. Vol II: Processing, U.S. Department of Agriculture, Agriculture Handbook No. 605

5.

Schnettker, T. (2006) Bewertung von Schleifmittel auf Unterlage für die Holzbearbeitung, Dissertation, TU Braunschweig

188


Surface Roughness of the Wood – Reality vs. Measured Figures Magoss, Endre and Tatai, Sándor University of West Hungary, Faculty of Wood Sciences, Sopron, HUNGARY Author for correspondence: [email protected]

ABSTRACT The standards are the pillows of the factory-independent, consumer oriented manufacturing. The international companies’ products sometimes come from different plants. The geographical distance could be more than a continent. The new generation of costumers uses more and more frequently the Internet to buy, without the physical contact with the offered goods. The mass production in the wood industry forces out small tolerant dimensions. The geometrical values are measured easily in general, whilst the quality of the surface is also important. This one is depends on some factors, so more complicated to gauge. The standardized, measurable quality figures help to keep to expectations of the costumers. The standardized figures applied to describe the metal surface roughness are found repeatable and useful. The classical calculation methods are not applicable for the surface of the timber or veneer covered materials. Special problems of choosing the measuring method, filtering or the figures repeatability are examined in this paper. The measurements of contact (stylus) instrument are compared with each other varying the material of the work piece and processing methods. The results of slice (2D) measurement are set against the topography of the wood.

INTRODUCTION Roughness characterises the fine irregularities on a machined surface. These irregularities can be determined by measuring the height, width and shapes of the peaks and valleys produced by wood working operations or by internal structural properties. The surface quality is a complex definition and it is characterised today by different parameters such as the more common Ra, Rz and Rmax parameters. Further details can be established using the Abbott-curve and its related parameters Rpk, Rk and Rvk. These parameters are standardised (DIN 4768 and 4776) and for their determination modern measuring units are commercially available. During the evaluation of the roughness due to woodworking operations we could observe that the surface of the given probe will have an even finish when optimal cutting parameters are applied. The anatomical roughness is located below the even surface, while the roughness due to woodworking operations is located above this surface line Fig1.

189


Fig. 1 To the calculation of roughness component due to woodworking operation Whether the wood species manufactured with non-optimal cutting parameters, the deformed surface produces more peaks in the measured contour. In this case the ratio of peaks to valleys more then anatomically determined. Moreover in the recent practice of high-speed machining, roughness due to machining is usually much less than the structure-based roughness, especially in the case of hardwood species with large vessels. The roughness due to machining usually depends on the following factors: cutting speed, chip thickness, processing direction relatively to the grain, rake angle of the tool, sharpness of the tool edge (tool edge radius) and vibration amplitude of the work piece. The commonly used figures of the surface roughness are very sensitive to the anatomical properties of the wood. The standard sometimes gives possibilities to evaluate the data, but to find the adequate parameters not easy. However an improperly chosen cut-off (length) for large-vessel wood species will produce artificial peaks in the profile, which also changes the values of roughness due to woodworking operation. (Fig. 2).

Fig. 2 The appearance of artificial peaks due to inappropriate cut-off [1] We needed special software to process the measurement data. The software has to be able to evaluate the measurement data in two different ways: on one hand it should apply a robust Gaussian regression filter in order to avoid (exclude) the artificial peaks in the area of the large vessels that are cut. On the other hand it should provide the option to manually remove the large vessels, which is required during the detailed evaluation of the roughness due to woodworking

190


operations.

EXPERIMENTS Our goal was to examine the correlation between the wood’s local anatomically determined texture and the measurable figures. The anatomy of the wood specimens differs by the kind of wood, and by the grooving conditions. However at this stage of our work we concentrated to the local differences of the surface, in consideration of reducing the factors effected on some anatomically different kind of species was chosen. Some is commonly used in Hungary and as well in Europe (Pine - Pinus sylvestris, Larch – Larix, Oak - Quercus patraea, Beech – Fagus sylvatica, Ash – Fraxinus, Aspen –Populus, Walnut - Juglans regia), some is exotic especially tropical (Balsa - Ochroma pyramidale, Etimoé - Copaifera salikounda, Kotibe - Nesogordonia papaverifera, Sipo – Entandrophragma utile, Sapelli Entandrophragma cylindricum, Okume - Aucomea klaineama, Zebrawood Microberlinia brazzavillensis). The tropical ones were chosen because of the special anatomy. The surface was planed by a Marunaka “Super Surfacer [2]”. This machine made a very good surface. The form of the chips give information about the machining process: if the chip is unbroken and smooth, the finished surface is even enough close to the optimal one. The moving gives a cycloid-free minimally waved base with no dust and chips in the vessels. The surface was scanned by a stylus type Perthometer. The radius of the tip was 2 micron. The Perthometer is 2D measuring equipment, it gives the heights of the surface by a plain, as of the Fig. 3.

Fig. 3 Components of the measured profile [3] The measuring by a line gives the cross section of the species. The evaluation of the measured data means to filter the roughness from the total profile, and generate characteristic figures. In the 3D measurement required to collect the topography of the surface. One possibility is to make a slices and from this ones generate the 3D model of the surface. Our method was to measure inside the square. The species were app. 50 mm x 50 mm with 20 to 25 mm height. The work piece was pressed to base surfaces as it seems on the Fig. 4. Each measuring line was parallel to other ones and to the edge of the piece. The moving was made by the precision screw of the coordinate table, the distance was 0.1 mm between the measuring lines. For the further evaluation

191


a program collected the data, 8000 pair of points by slice. The number of the slices was 175 for each piece.

Fig. 4 Two bases was installed on the coordinate table for the repeatability The time required for a topographical measurement almost a day for a piece, so we used this method for a reference. For a mass measurement required a calibration, we used an aluminium plate milled with 2 mm x 2 mm cross lines (Fig. 5.). The measured values help positioning a noncontact measuring equipments, and optical microscope or SEM.

Fig. 5 Aluminium calibrating piece

DISCUSSION We needed special software to process the measurement data. In the industrial world it is well known that linear filters are non robust, which means that any protruding peak or valley (artificial peak) leads to a distorted roughness topography and effects the calculation of surface parameters directly [4]. One solution is the robust Gaussian filter according to ISO 11562, by means of which you can „flatten out” the profile extension/distortion in the area of the vessels cut. The other possibility whether we want to evaluate the roughness due to woodworking operations of large-vessel wood species, it is reasonable to remove the vessels form the evaluated profile. The versions of the software „CurveCutter” were developed for the Department of Wood Technology and Department of Wood Machining at the University of West Hungary. The software has to be able to evaluate the measurement data in two different ways: on one hand it should apply a robust Gaussian regression filter or Csiha-method [5] in order to avoid (exclude) the artificial peaks in the area of the large vessels that are cut. On the other hand it should provide the option to manually remove the large vessels, which is required during the detailed evaluation of the roughness due to woodworking operations. In this case we can use the mouse and mark two points in the profile. The area between the two marked points (the vessel cut) will be removed and excluded from further evaluation.

192


The program is convenient for the calculation of the standardized parameters of surface roughness based on the profile P. It goes without saying that it also can display the material content curve (Abbott curve) Fig. 5 and calculate the corresponding roughness properties of P and R profiles in every case.

Fig. 5 Primary profile and the calculated Abbot curve The measured figures are in the figures following. 250

200

Pz: Pa:

150

Pt: Pmax: Pk: Mr1:

100

Mr2: Ppk: Pvk: 50

0_0 0_30 0_60 0_90 1_20 1_50 1_80 2_10 2_40 2_70 3_0 3_30 3_60 3_90 4_20 4_50 4_80 5_10 5_40 5_70 6_0 6_30 6_60 6_90 7_20 7_50 7_80 8_10 8_40 8_70 9_0 9_30 9_60 9_90 10_20 10_50 10_80 11_10 11_40 11_70 12_0 12_30 12_60 12_90 13_20 13_50 13_80 14_10 14_40 14_70 15_0 15_30 15_60 15_90 16_20 16_50 16_80 17_10 17_40

0

Fig. 6 Abbot parameters of the Balsa - Ochroma pyramidale vs. the position of slice 120

100 Pz: 80

Pa: Pt: Pmax:

60

Pk: Mr1: Mr2:

40

Ppk: Pvk: 20

0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 101103105107109111113115117119121123125127129131133135137139141143145147149151153155157159161163165167169171173175

Fig. 7 Abbot parameters of the Oak - Quercus patraea vs. the position of slice

193


The sample figures could be evaluated by the meaning of the Abbot parameters. Tab. 1 The average and the standard deviation of the Abbot parameters Beech

Balsa

Ochroma pyramidale Avg. Pz: Pa: Pt: Pmax: Pk: Mr1: Mr2: Ppk: Pvk:

74 6 129 120 18 7 84 21 34

St. Dev.: 23 1 46 44 1 1 2 4 15

Fagus sylvatica St. Avg. Dev. 47 9 5 1 68 15 63 13 17 6 6 3 78 4 17 4 23 6

Pine

Etimoe

Oak

Pinus sylvestris

Copaifera salikounda

Quercus patraea

Avg. 46 4 69 61 16 8 82 16 23

St. Dev. 4 0 10 10 1 1 2 2 3

Avg. 54 4 84 81 14 7 80 14 22

St. Dev. 8 1 12 12 2 2 2 2 3

Avg. 70 5 118 109 11 13 86 17 39

St. Dev. 13 1 20 20 1 2 2 2 10

Rk –depth of the so-called ‘core’ Rpk – reduced peak heights Rvk – reduced valley depths A1 – material ratio of peak elements A2 – surface of valleys Mr1 – ‘smaller’ material portion value of the core Mr2 – ‘higher’ material portion value of the core

Fig. 8 The meaning of the Abbott parameters The legend on the Fig. 8 calculated from the roughness profile meanwhile we had calculated from the primary profile. All the parameters are equals if we change the letter “R” representing the roughness profile to the “P” which means the figure of the primary profile. The filtering of the waviness is another possibility to add error into the evaluation process. The Abbot curve represents the effect of the manufacturing and the anatomical texture added. As in the samples it seems the peak elements represent fewer portions than the valley elements at small standard deviation. This means the parameters of the planning process was very close to the possible optimal ones.

CONCLUSION Our measurements proved that in most cases there is not enough to measure by a single line the surface roughness of wood pieces. The surface roughness parameters could be strongly vary by the changing the measuring line. The local surface parameter depends on the texture of the wood species. The small vessels not mean automatically smaller deviances, more important whether the free vessels are homogenously on the surface.

194


The highest relative deviances of Pa were measured for balsa and beach, oak and small deviances were detected for aspen, pine and larch. The Pk is depends on the position of the slice when the piece is beach, okume or zebrawood, the figure is not sensitive when we measure the balsa, pine, aspen or oak. Summarising the figures, we found that the local surface parameters measured not sensitive to the position of the slice when we examine especially aspen, pine, larch or kotibe. The evaluated parameters of other species were sensitive to the position of the measuring plane. In this case not enough one measurement to find the surface roughness. The cutting-of the vessels of wood by filtering or by hand is just one factor, generally need multiply measurement to evaluate the surface roughness of wood.

REFERENCES 1. Yuko Fujiwara, Yoshihisa Fujii, Shogo Okumura (2003) Robust Gaussian filter and threedimensional parameter to evaluate wood surface roughness as related to tactile roughness. pp. 359-365 In: Proceedings of 16th International Wood Machining Seminar 2. http://www.marunaka-jp.com/ENG_HP/Surfacer_oneway.htm 3. Horváth, S. (2008) A felületi hullámosság 2D-s és 3D-s jellemzése, a mőködési tulajdonságokra gyakorolt hatásának vizsgálata és elemzése PhD Thesys Zrínyi M. University, Budapest 4. Magoss, E. (2008) General Regularities of Wood Surface Roughness In: Acta Silvatica & Lignaria Hungarica Volume 4. pp. 81-93. 5. Csiha, Cs. - Krisch, J. (2000): Vessel filtration – a method for analysing wood surface roughness of large porous species, Drevarsky Vyskum 45(1): 13-22. 6. Sandak, J. – Negri, M. (2005) Wood surface roughness – what is it? In: Proceedings of 17th International Wood Machining Seminar 7. http://tropix.cirad.fr/index_gb.htm 8. Sinn, G. - Mayer, H. - Stanzl-Tschegg, S. (2005) Surface properties of wood and MDF after ultrasonic-assisted cutting, Journal of Materials Science 40: 4325-4332

195


196

Sanding - Surface Evaluation Poster Presentations


Study on Distribution and Elimination of Static Charge on Laminate Flooring Ma, Xiao

Zhou, Handong Ding, Tao

Zhou, Xihe Shen, Danni

Faculty of Wood Science and Technology, Nanjing Forestry University, 159 Longpan Rd., Nanjing, 210037, China ABSTRACT Fine wood dusts often attach to the surface of laminate flooring because of electrostatic force. Mechanical cleaning via roller brush can hardly get satisfying effect, making surface quality of the product affected. In this study, magnitude, polarity and distribution of static charge on laminate flooring were measured before and after pressing, cutting and profiling steps of the manufacturing process. The aim was to explore the reason and location of the generation of static charge. The results showed no definite distribution pattern of static charge on the flooring. The value of the charge ranged from around -1000V to 1000V. After treating the flooring by ion air gun or in high humid environment, the value was reduced significantly, indicating 2 possible dust detachment methods

INTRODUCTION Natural wood, a renewable resource as it is, is in short supply due to huge demand on it. An effective way to solve the problem is to adopt alternatives to natural wood products. With defect-free visual appearance and high erosion resistance, laminate flooring is treated as a promising alternative to solid wood flooring and has been used extensively. In 2010, sales volume of laminate flooring in China increased 12.3% compared to that of 2009, reaching 238 million m2, and accounted for 60% of the total flooring sales volume [1]. Although laminate flooring has been a popular product, many aspects are still to be improved in terms of product quality and manufacturing techniques [2]. Fine dust attachment to the surface of laminate flooring is a problem currently encountered by many laminate flooring makers. Ejection of goods sometimes occurs because of it. Many attempts have been tried, among which are equipping dust collectors or fans above the conveyor belt and manual dust removal. But none of them can get a satisfying result. It has been found that most dust attachment to laminate flooring is caused by electrostatic force. In this study, magnitude, polarity and distribution of static charge on laminate flooring were measured before and after pressing, cutting and profiling steps of the manufacturing process. Two static charge elimination methods were then evaluated. The aim was to explore the reason and location of the generation of static charge and lay a foundation of elimination of find wood dust attachment.

EXPERIMENTAL Measurement of static charge was performed in a laminate flooring production line. The

197


manufacturing process of the production line is as follows: Different layers of the flooring are stacked manually and pressed in a single layer press with temperatures reaching 190 degrees Celsius and up to 3MPa of pressure for 36 seconds. After pressed, the sheets are cooled to acclimate. Then, they are cut into planks by precision sizing saws. Finally, the boards are profiled by double-end tenoners. A motor driven roller brush has been mounted at the outlet of the double-end tenoner to remove dust on the flooring. The experiment was composed of 2 parts. In the first part, magnitude, polarity and distribution of electrostatic charges on laminate flooring were measured by a static electric meter (SIMCO FMX-003) before and after pressing, cutting and profiling steps of the manufacturing process. In the second part, two static charge elimination methods were evaluated. The first one applied an ion air gun ( Kunshan Enshuo Electrostatic Technology Co., AS-6202) to neutralize the static charges by blowing air with positive and negative charges. The second one was performed by increasing the conductivity of the flooring surface. An iron climate chamber with a linear size of 20cm was designed for the latter method. It was mounted at the outlet of a double-end tenoner. An ultrasonic humidifier was connected with the chamber by a hole drilled at the lower part of the chamber. Two slots were made at 2 opposing sides of the chamber so that the boards could pass through the chamber. As relative humidity in the chamber rose, a moisture film formed on the flooring surface, which increased the conductivity of the surface significantly. Meanwhile, large amounts of ions existed in the air because of high humidity. Both factors help diminish static charges on the flooring [3].

RESULTS AND DISCUSSION Static charge measurement 1) Static voltage before and after pressing Several boards were selected to measure the static voltage on the surface before and after pressing. The size of the board was 2.44m×1.22m. On every board, 72 evenly distributed measurement pointes were selected. The environment temperature was 30 ℃ and relative humidity was 44.8%. The measurement results showed that static voltage on the board was randomly distributed before pressing. Both positive and negative charges existed. The value varied from -800V to 400V. The static voltage decreased after pressing and tended to distribute evenly on the board, with the value between -20V and 50V. Fig.1 illustrates values from a sample board. The change of voltage distribution may be attributed to the following factors: After stacking, what on the top was the wear layer containing aluminum oxide and melamine resin. Both were poor electrical conductor. Large amount of static charges were attached on the wear layer after being produced. Measurement showed that the static voltage was as high as 3000V. The wear layer was wrapped before use. When part of it was taken out for stacking, stripping electrification occurred. Factors like the speed of unwrapping and the force applied would influence the static charges on the layer, causing the uneven distribution of static voltage. During pressing, the wear layer was firmly pressed by the steel sheet. Most of the charges were transferred, causing static voltage value to decrease after pressing and distribute evenly on the board.

198


Fig.1 Static voltage distribution before (left) and after (right) pressing 2) Static voltage before and after cutting Before cutting, the dimension of the board was 1260mm×1220mm. 72 evenly distributed points were chosen as measurement points. The board dimension was 415mm×1220mm after cutting. 24 evenly distributed points were chosen on the narrow board. The environment temperature was 29℃, and relative humidity 44.6% . According to the measurement results, electric charge distribution on the board was fairly even before cutting. All were positive with modest value variation between 0V and 60V. Little change occurred after cutting. However, voltage increase was observed along the kerf with the maximum value of 150V. Fig. 2 shows the voltage change on one sample board before and after cutting.

Fig. 2 Static voltage distribution before (left) and after (right) cutting Static charge change at this step may be caused by the following factors: After pressing, the boards were cooled to acclimate. During this period, static charges on the board contacted ions in the air, making the voltage value lower and distribution even. When subjected to cutting, high speed friction occurred between the board and the blade, increasing static charges along the kerf. 3) Static voltage before and after profiling

199


In this experiment, 3 positions along the production line were chosen to make the measurement: before profiling (position 1), after profiling but before the roller brush (position 2), after the roller brush (position 3). The board dimension was 1210mm×125mm. On every board, 20 evenly distributed points were chosen as measurement points. The environment temperature was 29.8 ℃ and the relative humidity was 42.2%. Measurement results showed that at position 1, distribution of static voltage was fairly even and all were positive charge with the value between 60V and 210V. The value rose at position 2, varying between 160V and 270V. The roller brush greatly changed the charge distribution. On most part of the board, voltage value increased around 400V, making it reach 390V-920V. The voltage change at the 3 positions was illustrated by fig. 3.

Fig. 3 Static voltage distribution at 3 positions of profiling step P1: position 1, P2: position 2, P3: position 3 The voltage increase at position 2 was mainly due to the interaction between the board and cutters or conveyor during profiling. Some dusts generated in the process were also charged, but most of them were sucked by hoods of the dust collecting system. The roller brush was designed to remove dust on the board. However, as the brush kept brushing the boards, static charges built up because of the friction between the brush and the boards, making the board more susceptible to dust attachment. Static charge elimination 1) Ion air gun The ion air gun was aimed at the position where the roller brush brushed the board. Air pressure of the gun was set at 50psi. The result is illustrated by fig. 4. The 3 positions were the same as what described in the previous experiment. It can be found that the voltage values at position 1 and position 2 are similar to the previous experiment, with voltage value ranging from 60V to 200V at position 1 and 180V from 350V at position 2. At position 3, a significant voltage decrease was found, with the value reduced to 40V-120V (temperature:28 ℃ , relative humidity:40.8%). The decreased static voltage, along with the air blown on the board, effectively lowered the amount of dusts attached to the board. Measurement on sample board showed a decrease from 530mg/m2 to 160mg/m2. 2) Relative humidity control The climate chamber was mounted at the outlet of the double-end tenoner, before the roller brush. The relative humidity was controlled between 60% and 80%. Results from sample board were shown by Fig. 5. The 3 positions were the same as the previous experiment. At position 1, the

200


static voltage was evenly distributed with the value between 180V and 320V. The voltage slightly increased at position 2 to 180V-350V. While at position 3, due to the influence of higher relative humidity, static voltage on the board decreased to 60V-90V (temperature: 26℃, relative humidity of the outside air: 41.8%). The amount of dust on the board also decreased significantly, from 150mg/m2 to 250mg/m2.

Fig. 4 Influence of ion air gun on static voltage on board at position 3 (left) Fig. 5 Influence of relative humidity on static voltage on board at position 3 (right)

CONCLUSIONS Distribution of static charge on laminate flooring was measured before and after pressing, cutting and profiling steps of the manufacturing process. Results showed that static charges were mainly induced by friction between the flooring and cutter or rubber roller of the processing machining. Stripping electrification also contributed to it, but was not the dominant factor. There was no definite charge distribution pattern on the flooring. The value of static voltage ranged greatly, from around -1000V to 1000V, causing enough adhesive force to fine dust. The static charge on laminate flooring could be eliminated by means of charge neutralization and increase of conductivity. Experiments showed that when the air pressure was 50psi or even higher, ion air gun could lower the static voltage on the board to as low as 120V. Dust on the sample board decreased from 530mg/m3 to 160mg/m3 as a result. When the relative humidity of air increased to 60~80%, the static voltage on the board could be lowered to around 100V. Dust on the sample board decreased from 510mg/m3 to 250mg/m3 correspondingly.

REFERENCES 1 China National Forest Product Industry Association. http://www.cnfloor.org/show.php?contentid=338 2 Gao, Zheng-zhong (2004). Manufacture and Utilization of Wooden Flooring, Chemical Industry Press, Beijing, China. 3 Lu, Cheng-zu (1991). Principle and Disaster prevention of Electrostatic, Tianjin Univercity Press, Tianjin, China.

201


Characteristics of Interior Décor of Hunting Lodge in Julin in Surface Quality Aspects Rozanska, Anna and Swaczyna, Irena Faculty of Wood Technology, Warsaw University of Life Science – SGGW, Nowoursynowska 159, 02-776 Warsaw, e-mail: [email protected] ABSTRACT The hunting lodge in Julin is a mansion build of wood which in the years 1880-1944 was a property of Roman and Alfred Potocki from Lancut (http://www.zamek-lancut.pl/). It is one of the most beautiful aristocratic residences in Poland. The grandeur and standard of the mansion aroused admiration among the people of that time. The building is one of the few ones remaining until today in its original form. Apart from the layout of the interiors, also some fixed elements of wooden furnishing were preserved: wooden windows and doors, floors, panelling, ceilings, stairs and builtin furniture. The paper consists of a historical description of the monument and the characteristics of certain interiors. The wooden historical building was examined within the scope of the quality of the internal wooden surfaces: staircase, panelling, doors, windows, parquet and furniture. The quality of the surface was evaluated as to its aesthetic and constructional properties, as well as to its durability and preserved condition. The surface quality analysis was carried out through visual inspection of the structure and also through chemical analysis of the wood finishing materials on the surfaces of the elements of interior furnishing. The photographic documentation and state of the art of the ones were acquired.

a

b

Photo 1. Hunting Lodge in Julin; view from South-West (a) and view from the Eastern side (b)

INTRODUCTION The Hunting Lodge in Julin (Photo 1a,b), was built during just one year in 1880. It was constructed on a layout resembling an H-shaped plan, although it seems rather a picturesque assemblage of several independent forms. A part of it has basement underneath (with a brick foundation) and its walls, with horizontal timber construction, are made of larch logs. On the outer side they were supposed to be covered with a layer of varnish in order to strengthen them [1]. The monograph of this object mentions the logs’ square section with rounded corners, which makes them resemble round logs [2]. The roof was covered with wooden shingles (shakes). 202


The designer and construction supervisor was W. Pannenka, probably of Czech origin, who had been working for the Potocki family for years. According to the signature, the roof truss was constructed by Stanisław Podgórski from Leżajsk. Other carpenter works were entrusted to a Wawrzkiewicz from Łańcut. The local craftsmen took a significant part in the furnishment of the interiors, although most of it was imported from Paris and Vienna [3]. The Hunting Lodge in Julin has been preserved until today in its original state, yet without the internal movable furnishings. Apart from the interior’s layout, some permanent elements of the furnishings were preserved: wooden windows and doors, floors, panelling, ceilings, stairs, book cabinets, stoves, coat rack in the Corridor or a hanging lamp in the Hall on the ground floor. The Hunting Lodge in Julin is an eclectic building. Its form is a good example of the Alpine style, which was very popular in Europe in the 19th century. The purpose of the building influences the characteristics of its details, associated with the architecture of hunting lodges. Some experts have noticed neogothic elements in the details, like for instance the tracery decor of the beacon with quatrefoil and rhomb motifs, or the chimneys made of red brick. Window frames and woodworks, the roof tops and the awnings are decorated with openwork wood-carving ornaments typical for the architecture of health resorts of the second half of the 19th century (Photo 3). Julin is also often compared with the hunting lodge in Niemodlin and with the project of a villa from the Paris Exposition of 1867 [4], as well as with prince Antoni Radziwiłł's Antonin lodge in Poznańskie district dating to 1820s or Promnice near Pszczyna lodge from 1868 belonging to Jan Henryk von Pless from Książ, or Spała in Masovia from 1884 belonging to the tsars Alexander III and Nicholas II. DOORS, WINDOWS, STAIRS, PANELLING AND CEILINGS The interiors have eclectic character as well. The central axis of the building is set by the double, two-winged door (which can be opened to the inside and to the outside) and by the windows placed on both sides of the door and separated from it only with poles.

Photo 2. Projection of the Julin Lodge's ground floor after the extension in 1927 (Cholewianka-Kruszyńska 2003)

Photo 3. Part of the Hall with stairs

The entrance leads to the Hall, occupying the whole width of the risalit, with a representative staircase with two flights of stairs. In the Hall, on the building’s axis, under the landing of the stairs, there is a passage to a narrow, transverse Corridor. On the Western side of the Hall, there is the 203


Living Room, and on the Eastern side, the Dining Room. Both of them can be entered either from the representative part, or from the corridor for the staff. The shape of the Living Room, almost square and clear span, is broken by a rectangular risalit lit by a wide window. The Dining Room, shaped analogically to the Living Room, is smaller than the latter because of two small utility rooms: a smaller, dark square one, which can be accessed from the Corridor, and another, rectangular one, having a narrow window, that only can be accessed from the Dining Room. Those rooms, together with a part of the Corridor, create two symmetrical recesses, limited by two pillars, in the surface of the Dining Room wall. Those pillars sustain the ceiling (Photo 2). The stairs leading from the Hall to the first floor, consist of joint four flight landing stairways. Both symmetrical stairways, that at the base are parallel to the main axis of the building, turn at right angle after square landings and meet on the middle landing. Afterwards, they split once more towards the sides, turn at right angle and end up on both sides of the Hall on the first floor.

Photo 4. Fragment of the Hall ceiling with the rosette and antique lamp visible Photo 5. Double-winged door from the side of the Hall

The walls of the Hall were covered with pine panelling on the whole height, in a tectonic composition of a recessed panel pedestal with a baseboard, of vertical butt joint boards (each about 200mm wide), and of horizontal entablature with a wide frieze decorated with demi-columns of trilateral transverse cross-section (Photo 3). They are analogous to two octagonal wooden pillars with the cross-section of 130mm (with a central annulet and a ring-shaped top end instead of a capital) that sustain the middle landing of the oak staircase. Another elements which have octagonal cross-section are the posts which sustain the handrail. Metal elements with marvellous openwork have been used as balusters. They either have a highly stylised floral pattern à la art nouveau or are composed of modernised rocailles, making reference to the outer woodwork decor. The strings of the stairs are topped with capping made of profiled batten. The wooden coffered ceiling, made of thick profiled pine beams, is decorated in the middle with an octagonal rosette, from which a lamp is hanging (Photo 4). Between the staircases there used to be a billiard table, lit with double chandelier, connected to electricity supply at a still existing point below the cartouche of the coat of arms. The lighting of the interior is completed with numerous lateral sconce lamps. The walls were decorated with hunting trophies and the coat of arms of Pilawa, with a painted shield and with the panoply represented on a label within a woodwork ornamented decoration, was fixed to the bottom part of the higher flight of stairs. The oak double-winged paneled doors placed symmetrically on the Eastern and Western walls of the Hall, were ornamented with pine pediments with wood-carving decorations of pinnacles and openwork top ornaments (Photo 5). The height of the doors is adjusted to the panelling and amounts to 3565mm. The wings are 2675mm high and the door is 1815mm wide. These doors are identical to the entrance door in the Dining Room from the side of the Eastern porch. The window openings in both lateral rooms were given analogous pediments. The lower part of the windows is coupled and of a type typical in Poland at that time (double window 204


with one part opening to the inside and another to the outside, mounted on a single wooden frame), they are 1100mm high and 1350mm wide along the internal line, and above there is an extension consisting of a horizontal pivoting window. The total height of the window along the internal line amounts to 1750mm.

a

b

c

Photo 6. Fragments of the Living Room panelling with the "Pilawa" coat of arms on the Southern wall (a,b) and on the Northern wall (c)

Probably also the walls of the lateral Living Room were covered entirely with pine panelling with a composition similar to that of the Hall (Photo 6a,b). Apart from the recessed panel pedestal and the vertical butt joint boards, it was topped with a frieze with neogothic arcading with the motif of trifoil in the archivolt, under a richly profiled cornice moulding. Pilasters, topped with angulated cornice moulding and ornamented with mascarons below the cornice, were placed over the panelling. In the pedestal part, the recesses beneath the windows were decorated with a relief of the Pilawa coat of arms with stylized floral decor on the sides (Photo 6c), analogous to the coat of arms placed above the window in the Western risalit. Only a fragment of the Living Room panelling has been preserved until today, between the Southern windows and in the risalit. Inside the risalit, probably intended for a writing desk, a large glassed panel was designed, 2300mm high and 1760mm wide, starting 340mm above the floor level. In the 1920s outer wooden blinds were installed.

Photo 7. The Dining Room panelling

Photo 8. Intertwined veneer panelling

The panelling of the Dining Room was limited only to the recessed pedestal (Photo 7). The ceilings of the Living Room and of the Dining Room are coffered, made of thick and richly profiled beams, with a frame ornament of battens. They were crafted of pine wood. The beams have greenish paint graining, which creates an interesting colour effect together with the orange hue of the battens and of the coffers’ internal elements. The ceiling in the Dining Room, within the narrow part on the Eastern and Western side, is sustained by two wooden pillars placed in the corners, having square transverse cross-section of 150 x 150mm, with a part of the pedestal marked with battens, its size being 240 x 240mm. The pillars also have a lowered pseudo-capital cornice moulding, above which there is a frieze marked with openwork wood-carved corners, imitating corbels. Observing it with scrutiny one may notice, among their stylised ornaments, the “RP” monogram marked out with dark paint. Twin quarter turn staircases lead from the Hall to the first floor, which has a clearly less representative and more residential character. The railing pattern on the first floor was changed into

205


a more humble one, with an openwork motif of circles and quatrefoils elaborated of wooden battens. The walls in the hall on the first floor are covered up to one-third of their height (panelling height: 1030mm) with interesting panelling with intertwined pattern, like that of a basket, elaborated of 30mm wide strips of pine veneer interlacing at right angle. The same type of panelling repeats in the Corridor on the ground floor (Photo 8). From the Hall on the first floor, one could access six rooms. According to the information included in the monograph, the ceilings of the rooms were finished with stucco work, and their names were associated with the colours of the wall tapestry materials – e.g. blue or pink.

a

b

Photo 9. Fragments of the Library ceiling with star grained (a) and wooden (b)

In the one-storey part of the building, after the year 1883, built-in oak cabinets were installed for the purpose of the Library. Each cabinet has a wood-carved label on the top, indicating the section of the book collection it was containing. The ceiling of the Library is very interesting, with its spruce graining and with battens placed in a frame pattern with stars in the corners. The central bigger star is painted in an illusion-like manner (Photo 9a,b). FLOORS The floors were probably also elaborated by local craftsmen. The floors of the three representative rooms on the ground floor were composed of separate basic square panels with geometrical patterns, and according to the archives they were imported from Lviv. The origin of the floor patterns was probably the same, and it was different for each of the rooms. The contrastive and distinctive floor of the Living Room was made of oak and hornbeam (Photo 10a). The floor in the Hall, made of oak and ash with similar hues, makes use of a highly decorative effect of different light reflection by elements whose fibres are placed at different angles towards those of the adjacent ones (Photo 10b). The same effect may be observed on the Dining Room floor, also made of oak and ash (Photo 10c). The above-mentioned floors were elaborated of solid elements connected with spline joints, 4,5mm thick, that enter the groove 18mm deep. The usable wear surface on the face side is right now 5,5mm thick. The floors of the less representative rooms (the Corridor on the ground floor, the Hall on the first floor, etc.) were made of oak parquet with the herringbone pattern. In the rooms on the first floor only the subfloor made of boards has been preserved. The rooms created during the subsequent extensions and reconstructions of the building were given only simple herringbone parquets. Only in the count Alfred Potocki’s Northern Living Room built in the 1920s and in the Guest Room located directly above, the parquet has been laid in an interesting manner, forming a pattern of four square elements joined with others at right angle, with square insets in the corners. All the baseboards were profiled with cuttings and central cove mouldings. The floors in Julin differ considerably from the decorative floors, made probably in 1830s, of the Łańcut Castle [5], which at that time was also owned by Roman Potocki.

206


a

b

c

Photo 10. Decorative wooden floor pannels in the Living Room (a), Hall (b) and Dining Room (c)

ACKNOWLEDGMENT The project forms part of a research grant number 1374/B/P01/2008/35 regarding the analysis of design, construction, and production technology of decorative wooden floor of manors and palaces in south-eastern Poland. REFERENCES 1.

Łowiec 1881 (11): 12

2.

Cholewianka-Kruszyńska, A., (2003) „Hunting Lodge in Julin, of Roman and Alfred Potocki from Łańcut” (Julin Pałacyk Myśliwski Romana i Alfreda Potockich z Łańcuta) 1880-1944, {in:] The Art of Hunting: Intellectual Contexts. Court Hunting. Papers of the 2nd Session organised by the Zamoyski Museum in Kozłówka, 10-11th May 2003, Zamoyski Museum in Kozłówka, pp.100-141.

3.

„Letter from Roman Potocki to Józef Michalski, 17th November 1879” (List Romana Potockiego do Józefa Michalskiego z 17 listopada 1879r.), AGAD, APŁ, sygn. akt 3794.

4.

Błotko, V.,Błotko, A., (2002) „Hunting Lodge in Julin. Contribution to the object’s monograph” (Pałacyk myśliwski w Julinie. Przyczynek do monografii zespołu), [in:] Rzeszowska Teka Konserwatorska, Rzeszów.

5.

Kossakowska-Szanajca, Z., Majewska-Maszkowska, B., (1964) Łańcut Castle (Zamek w Łańcucie), Arkady, Warszawa.

207


Influence of Climate on Surface Quality of Antique Wooden Flooring in Manor House Rozanska, Anna and Tomusiak, Andrzej and Beer, Piotr* Faculty of Wood Technology, Warsaw University of Life Sciences, Nowoursynowska 159, 02-776 Warszawa, e-mail: [email protected]

ABSTRACT The basic requirement for wooden floors is that they should be durable. The wood of the floor is exposed to various factors causing its gradual wear. Among them there are biological (bacteria, algae, lichens, insects and others), chemical (acids, bases, salts, and aerosols), physical and chemical (light, radiation, high temperature, fire) and physical and mechanical factors (low temperature, changes in humidity, mechanical forces). The surface quality depends on the features of the material from which it was manufactured, especially its hardness, flexibility, strength properties, wear resistance and resistance to microorganisms. Therefore, the quality of the floor depends on the kind of wood (species, structure, chemical composition) and on the environment in which it is placed. The article presents an analysis of the climate of manor house interiors in relation to the external climate, the manner of heating the room, number and activities of the people residing inside and the hygroscopic materials present there. The basic manor house climate parameters taken into account were: temperature, humidity and the speed of air flow, as well as their changes over the buildings’ history. The wood ageing factors whose influence was analysed, were: low and high temperatures, frequent changes in humidity, UV radiation, static and dynamic loads (stress) and the influence of microorganisms on the flooring surface quality. Similarly, the importance of the floor’s structure was analysed, together with the manner and quality of its installation, as well as the usage conditions and the maintenance measures. Floor surface quality was examined by evaluating its macro and microscopic structure and by taking samples for chemical and strength tests.

INTRODUCTION Flooring is an architectural element that, together with walls and ceilings, constitutes an integral part of the interior. It is a valuable element of antique interiors. Depending on the materials used, floorings can be divided into: made of stone, ceramic, plastic and wooden. Wooden floors can be made of planks, wooden blocks, panelling, wooden mosaic, laminated boards, solid wood boards and other engineered wood products [1]. Wooden flooring is a set of individual wooden elements placed on an underlying support structure or a subfloor. It is made of modules of flooring boards, i.e. prefabricated units composed of elements fixed together by means of a system of joints. They can be solid wood boards or laminated boards, placed usually on a continuous support structure [2].

208


a

b

Illustration 1. View of the Tarnowiec manor house (a) and its layout with numbers of the rooms where wooden flooring was preserved.

The flooring chosen for the analysis consists of antique floors from the Tarnowiec (Ill. 1a,b) manor house. It was built in the 1830s, in the style of late Polish Classicism. It includes a preserved set of oak tile floors within the gardenside rooms. The history of the building, as well as its design, structure and state of preservation of the flooring (among others, traces of insect infestation on fragments of the flooring in a room that after the war was used as a grain depot for the municipal cooperative) show that the floors date back to the 1830s (Ill. 2a, b).

a

b

Illustration 2. Flooring patterns of the Rooms no 1, 2,4 and 6 (a) and Rooms no 5 (b).

The basic requirement for wooden floors is that they should be durable. The wood of the floor is exposed to various factors causing its gradual wear. Among them are biological (bacteria, algae, lichens, insects and others), chemical (acids, bases, salts, and aerosols), physical and chemical (light, radiation, high temperature, fire) and physical and mechanical factors (low temperatures, changes in humidity, mechanical forces). The surface quality depends on the features of the material from which it was manufactured, especially its hardness, flexibility, strength properties, wear propensity and resistance to microorganisms. Therefore, the quality of the floor depends on the kind of wood (species, structure, chemical composition) and on the environment in which it is placed. INFLUENCE OF MATERIAL ON ANTIQUE WOODEN FLOORING DURABILITY Floors are made preferably of older trees of high density, heartwood, with extractive compounds (resins, essential oils, tannins, fats). Technically, the best kind of wood is the oak ring-porous wood, taken from the heartwood. This is the kind of wood which was used for the flooring in the Tarnowiec manor house. It is durable: hard, resistant to abrasion, flexible and contains tannins that make it more resistant to microbiological factors. It is easily processed

209


and smooth surface can be obtained. The growing percentage share of summerwood is positively correlated with wood’s durability, hardness, shrinkage and resistance to abrasion. The share of summerwood usually tends to grow together with the width of growth rings, for this reason the growth rings are a reference indicator of the technical properties of oak wood [3]. The wood used to manufacture the Tarnowiec manor house floorings was wood of medium growth (3-4 rings for each 1 cm of the radius) and wood of slow growth (>4 rings for each 1 cm of the radius). The wood of medium growth has uniform structure, so the wear of the flooring surface was more evenly distributed. The best wood is the one that is straightgrained, because it helps to avoid spallings and surface roughness during the profile cutting. Sapwood is inadmissible on the front surface of the flooring elements. The most valuable kind of wood is oak heartwood from the radial section with large pith rays (lustre) [4]. In the Tarnowiec mansion, oak wood of mixed radial and tangential section was used. Taking into account that the radial section is twice as resistant to wear as tangential section, uneven floor level is currently visible.

a

b

Illustration 3. Scheme of floor structure in case of subfloor boarding, Room no 1, 2 and 6 (a) and in case of flooring placed directly on the beams, Rooms no 4 and 5 (b).

The floorings of the antique Tarnowiec manor house are fixed with forged nails to the subfloor (32 mm thick) which in turn is nailed to beams placed directly on the ground, whose dimensions are: 160 x 90 mm (Ill. 3a, b). The distance between the beams amounts to 535 mm, 600 mm or 720 mm, depending on the kind of tiles, while the recommended range is of 650-700mm for flooring that is 25mm thick and 800-900mm for flooring which is 32mm thick [3]. The flooring in two rooms of the Tarnowiec manor house is placed directly on the beams, with no boarding. The tile dimensions vary from room to room and amount to: 645 x 645mm in Room 1 and 2, 580 x 580mm in Room 3 and 4; and 765 x 765mm in Room 5. They are 29mm thick except for Room 5: 32mm thick. The tiles are assembled using spline joints. The tiles have single-layer structure, and their wear surface amounts to 12mm. The individual elements within the tiles are fixed using tongue and groove joints. Alongside the walls a wedge-shaped distance of 15-20mm was left, which is covered by battens and socles. INFLUENCE OF SURFACE FINISHING ON ANTIQUE WOODEN FLOORING DURABILITY In the Tarnowiec manor house, the flooring was placed evenly and was well levelled. The individual flooring elements initially did not have any uneven points, cracks, fissures, indentations nor spallings that are visible nowadays. After the floors were placed, they must have been sanded with a hand scraper along the fibres. Before manual smoothing, the floor was partially rinsed with hot water and after the smoothing it was polished with steel shavings [3].

210


The antique wooden floorings of Tarnowiec were not soaked with substances that would significantly decrease the wood sorption. However, they were protected against it, as well as against dirt penetration and forming of stains, by means of waxing or varnishing (repetitive soaking with hot varnish until total saturation of the wood’s surface was achieved). It is a time-consuming process, which results in the darkening of the wood colour and dimming of its lustre, yet it provides a durable flooring surface finishing effect. There were no visible scratches on the surface and it could be maintained in cleanliness easily. In case of flooring surface renovation, the wax polish was removed with gasoline, and varnish - with dissolved alkali (water solution of sodium hydroxide). After the war, polishes containing wax, paraffin and volatile organic solvents (turpentine, gasoline or BTEX) started to be used massively for the purpose of wooden floor maintenance. The floor finishing influences its durability – it increases the abrasion resistance. INFLUENCE OF INTERIOR FLOORING DURABILITY

MICROCLIMATE

ON

ANTIQUE

WOODEN

The climate of the Tarnowiec manor house interiors depended on the external climate, the manner of heating the room, the number and activity of the people residing inside and the hygroscopic materials present there – the plaster on the walls, the floorings and the furnishings. The basic parameters of the climate of the interiors were: temperature, humidity and air flow speed. Spacious and high rooms of the 19th century brick manor house in Tarnowiec were heated with stoves and were permanently underheated. Windows and doors were not well sealed and the thermal insulation of floors and ceilings was inefficient. It is possible that a part of the representative rooms used to be closed for the winter period and that only the smaller living room was heated for the needs of the household. The hall was heated only with a fireplace, although it was situated in the middle of the mansion. The temperature of the interiors was about 10°C lower than at present (in the 19th century the recommended temperature range was 15-20°C). The temperature would be lower close to the walls and close to the floors of the rooms on the ground floor, while excessive air flows caused additional drops in temperature. The relative humidity of the air must have also been high; right now, in the manor house in Tarnowiec, which is not being used but is protected (with no additional humidity sources, with renovated roof and rain gutter system), according to the tests carried out, it amounts to 75-95%. Comfortable relative air humidity, at a temperature of 22°C in winter, amounts to 40-45%, and in summer: between 40% and 60% [5]. Initially, the Tarnowiec manor house (similar to other manor houses) had neither a kitchen nor a water and sewage system. Water for washing up, cooking and potentially also washing clothes and dishes (provided that it did not take place in separate buildings) was brought in buckets, which caused sharp local increase in relative air humidity. The manor house in Tarnowiec, throughout its history, would alternately fall into ruin and then be renovated. Leaking roofs (mostly next to the chimneys) caused that the water originating from rainfall or melting snow to pour onto the floor. When temperature fell below zero in rooms where the floor was covered with water, the unbound water present in the wood of the floorings would freeze. Moreover, excessive rate of heat penetration through the external walls of the manor house caused the vapour present in the air to liquefy on the walls

211


at the dew point and caused it to move further into the building. Fungi microorganisms would develop on humid surfaces, attacking also the wood of the flooring. Heating with stoves caused local overheating of the flooring situated next to the stove with simultaneous underheating of the parts located at the other end of the room. Stoves caused abrupt changes in temperature and relative air humidity. The sun shining through large windows also caused local overheating of the flooring and exposed it to the impact of UV radiation. During the summer period, western windows would cause increase in temperature inside the room. In the manor house in Tarnowiec, the temperature in unheated interiors depends on the external temperature. Nowadays, the temperature inside the unused manor house is 4°C higher than outside (sun operating and the number and activity of the people residing inside). Due to the fact that the partial pressure of water vapour within the interiors is higher than outside, the absolute humidity in the interiors is also higher than outside [6]. In practice, the relative humidity in the interiors is up to 10% higher than in theory, also because of the presence of people and hygroscopic materials in the rooms, as well as incomplete air exchange (mixture of external and internal air). Currently, the conditions of relative air humidity inside the Tarnowiec manor house stand at 95%, with the wood moisture equivalent of 15,7% for elements of flooring placed on boarding and 35,6% for elements of flooring placed directly on the beams and attack of fungi. Central heating is to be installed there this year, causing an abrupt reduction in relative air humidity during winter and, as a result, deterioration of the wooden flooring condition [7, 8, 9]. Frequent and rapid changes in relative humidity of the interior’s climate after switching the central heating on or off (May and October) are most harmful for wood. The relative humidity in the winter period is too low (φ=23%), while in the summer period it is too high (φ=76,5%), rising up to 90% (due to the ventilation of the room with humid air from the outside) [5]. FACTORS INFLUENCING WOOD SURFACE QUALITY IN MANOR HOUSE INTERIORS In closed interiors of the manor house in Tarnowiec, apart from the impact of water, changing temperatures, radiation and stresses [10], the flooring has been additionally exposed to water coming from the leaking roof, humidity penetrating through the walls due to lack of rain gutter system and humidity absorbed from the ground during the period when rainwater was not drained away to a sufficient distance from the mansion walls. The flooring of the given rooms of the manor house, depending on the layers of its structure, is in very different condition, which translates directly to their state of preservation. The manor house in Tarnowiec is partially cellared. In the places where there is no cellar, the subfloor beams are placed in a layer of sand or clay that was laid directly on the ground. In the cellared areas, they are placed in a layer of sand or clay filling the surface irregularities on top of the cellar ceiling. In such cases, the layers are much better protected against the penetration of groundwater and, moreover, they are ventilated with the air filling the cellars,

212


a

c

b

d

Illustration 4. Microscopic image of the front (a,b) and bottom (c,d) surface of an element of flooring placed on boarding, Room no 1.

a

c

b

d

Illustration 5. Microscopic image of the front (a) and bottom (b) surface of an element of flooring placed directly on the beams, Room no 5

213


which always have ventilation openings. The tiles, as explained above, are usually placed on a subfloor, but the tile flooring of two rooms is placed directly on the beams, with no boarding. The beams are placed in parallel to the shorter sides of the rooms, except for the abovementioned Tarnowiec interiors, where they are placed diagonally at 45 degrees angle. As a result of the differences of beam layout in Tarnowiec, the strips of tiles are placed there diagonally. In each case, there is a layer of sand between the boards of the subfloor and the tiles, whose function is to fill in the irregularities of the bottom sides of the tiles, finished with the use of plane, and to level the face side of the tiles. The usage of flooring causes the sand to move, crush and migrate through the gaps in the joints onto the front side of the tiles, resulting in their faster wear. The evaluation of the Tarnowiec manor house flooring was carried out visually (macro and microscopically) and by examining its roughness, hardness, wear propensity and resistance to scratches. The test samples were taken from flooring tiles placed on the boarding and beams in the rooms situated over the cellar, as well as from elements placed directly on the beams in the uncellared rooms of the manor house. The results of macroscopic observation showed that the wood of the flooring placed on the boarding is well preserved, with cohesive structure and nice, dark colour with no major discolorations, scratches or indentations (Ill. 4a, b). The utilisation of flooring placed directly on the beams in rooms 3 and 4 (with no boarding), caused indentations to appear on the front side of the elements in the proximity of joints, due to delaminations alongside the fibres formed as a results of the stresses caused by utilisation. The structure of those elements was less coherent, the wood had a greyish colour, and in some places signs of insect and fungi infestation were visible (Ill. 5a, b). Some elements were attacked by fungi only on the bottom surface, whose structure was degraded. The elements damaged by fungi had spongy structure, dark brown colour (brown-rot), were very light and developed cracks after drying. The roughness of the front surface of the flooring placed on the boarding was relatively low, below 0,152mm in case of tangential section and 0,098mm for radial section. Due to better abrasion resistance of the lustre (pith rays) in the elements manufactured of radial section, whose lustre was visible, the roughness was also clearly visible macroscopically and amounted to 0,5mm. The roughness of the front surface of the flooring with no boarding was even up to 0,278mm. The surface of the bottom side of the flooring tiles was less rough, showing clear traces of saw, which was used for its processing. Wood hardness tests in accordance to the Brinell method were carried out in compliance with the PN-EN 1534:2002 standard [11], by indenting a 10 mm diameter ball into the wood with the force of 1 kN, growing during 15 sec and maintained during a dwell time of 25 sec. The hardness of the wooden flooring placed on the subfloor boarding amounted to 38,97[ N / mm 2 ] for the radial section and 43,94[ N / mm 2 ] for the tangential section. The hardness of the wood with no boarding was significantly lower: 25,66[ N / mm 2 ] for mixed section. In case of the elements affected by microbial infestation, it was impossible to perform the tests in compliance with the standard due to their insufficient hardness. For comparison, the hardness of contemporary oak flooring elements was tested, which amounted to 44,39[ N / mm 2 ] in laboratory conditions (Table 1).

The wear resistance of wooden oak flooring dating back to the 1830s was measured with the Taber method, based on the loss of thickness and loss of mass, in accordance with the PN-EN ISO 5470-1:2001 standard [12]. The average loss of thickness of oak wood amounts to 0,18mm (with standard deviation of 0,03mm and coefficient of variation 14%), while the

214


average loss of mass - 0,204g (with standard deviation of 0,036g and coefficient of variation 17,5%) [13]. Table 1. Results of Brinell hardness tests the wooden flooring placed on the subfloor boardingthe radial section Hardness [N/mm2]

38,97

the wooden flooring placed on the subfloor boardingthe tangential section 43,94

Standard deviation [N/mm2]

1,48

5,92

4,07

3,65

Coefficient variation [%]

3,79

13,48

15,86

8,22

35,98

31,98

17,43

37,02

of

Characteristic value [N/mm2]

the wooden flooring with no boarding

contemporary oak flooring elements

25,66

44,39

The most harmful factor for wooden flooring surface quality in Tarnowiec are the frequent variations in humidity (constant wood moisture of up to 10% ensures high wood durability, e.g. the wood of Egyptian Sarcophagi [15, 5]). Frequent changes in wood moisture cause the shrinkage and swelling phenomenon. According to some researchers, wood damage takes place during this process because the changes primarily affect the external layers of wood [16]; according to others, it happens as a result of the sorption stresses during rapid adsorption and desorption processes [17]. The processes taking place within the wood during the changes in relative air humidity cause the formation of cracks or cause the floor to rise when its moisture is higher (shrinkage coefficient, swelling coefficient). For this reason, the choice of the wood's section is more important than its species. The swelling coefficient alongside the direction of fibres is much lower than the swelling coefficient in the radial direction, which in turn is lower than in case of the tangential direction. The coefficient of shrinkage uniformity is also important – it is the ratio between the tangential shrinkage coefficient value to the radial coefficient value [3]. Influence of low temperatures is also harmful. In those rooms of the Tarnowiec manor house that were not heated during winter, the unbound water in cell cavities used to freeze. This process caused its volume to increase by 9% and caused it to evaporate from the cell walls towards the forming ice [18, 19, 20, 5]. This resulted in the apparition of microcracks having the characteristics of desorption cracks (while drying up, the cell wall is stretched by the ice). It did not reduce the resistance to compression [21, 22]. Influence of high temperatures, which took place in the manor house in the immediate proximity of stoves and as a result of the action of the sun, caused changes in the chemical properties of hemicellulose, cellulose and lignin [23, 24].

As a result of the impact of ultraviolet radiation, the wood surface suffered decomposition (UV rays penetrating up to 5mm deep [25, 26]). Due to photochemical degradation, the surface of the wood darkened. In deeper layers, cellulose oxidised [27, 5].

215


Static and dynamic stresses below the endurance limit, associated with the usage of floorings, caused elastic and plastic strains of the wood and influenced the internal strength distribution within the wood. They could enter in interaction with the desorption stresses, intensifying deformation and apparition of cracks. Repetitive stress cycles resulted in material fatigue, damaging its structure [5].

However, the greatest damage to the antique floorings of the Tarnowiec manor house was caused by microbiological corrosion. The beams, the subfloor and the wooden floors were exposed to the attack of fungi and other xylophagous organisms. The floorings placed directly on the ground were attacked by the common furniture beetle (Anobium punctatum De Geer), and the infestation was most intense in the most humid room, which after the war served as a grain depot of the municipal cooperative. The floorings of the Tarnowiec manor house are infested with house fungi, in spite of the fact that oak, due to its chemical composition, is particularly resistant to them. The fungi, in order to develop, need high relative wood moisture, which is inevitable in case of the beams laying directly on the ground and the lack of ventilation between them. ACKNOWLEDGMENT

The project forms part of a research grant number 1374/B/P01/2008/35 regarding the analysis of design, construction, and production technology of decorative wooden floor of manors and palaces in south-eastern Poland. REFERENCES

1.

Turant, J. (2004) Posadzki, maszynopis [Flooring, typescript].

2.

PN-EN 13756: 2004. Wood Flooring.

3.

Korzeniowski, A. (1956) Posadzki drewniane [Wooden flooring], Warszawa.

4.

Woźniak, A. (2005) Charakterystyka posadzek drewnianych w Muzeum Zamoyskich w Kozłówce, praca magisterska, Wydział Technologii Drewna SGGW [Wooden Flooring Characteristics in the Zamoyski Museum in Kozłówka, M.S. thesis, Faculty of Wood Technology, Warsaw University of Life Sciences], Warszawa.

5.

Kozakiewicz, P., Matejak, M. (2000) Klimat a drewno zabytkowe [Climate and Antique Wood], Warszawa.

6.

Krzysik, F., Sobczak, K. (1960) Wilgotność drewna w pomieszczeniach ogrzewanych centralnie [Wood Moisture in Interiors with Central Heating], Sylwan 9: 29-43.

7.

Bieńkowski, T. (1995) Całoroczne badania zmian wilgotności drewna w pomieszczeniach muzealnych i kościołach, praca magisterska, Wydział Technologii Drewna SGGW [Wood Moisture Changes over the Year in Museum and Church Interiors, M.S. thesis, Faculty of Wood Technology, Warsaw University of Life Sciences], Warszawa.

8.

Schotes, P. (1973) Richlinien für die Beheizung von Kirchen, Das Münster (26) 1-2.

9.

Stadtmüler, P.A. (1972) Orgel und Heizung, Das Münster (25) 4.

10.

Monck, W. (1980) Notwendigkeit und Moglichkeit des bautechnischen Holzschutzes, Holzindustrie 5:3-6.

216


11.

PN-EN 1534:2002. Wood flooring- Determination of resistance to indentation- Test method.

12.

PN-EN ISO 5470-1:2001 Rubber- or plastics-coated fabricks- Determination of abrasion resistance- Part 1: Taber abrader (ISO 54-1:1999)

13.

Swaczyna, I., Tomasiak, A., Kędzierski, A., Koryciński, W., Policińska-Serwa, A. (2009) Indentation and abrasian resistance of decorative wood flooring of the Castle in Łańcut [w:] Ann.WULS-SGGW, For. And Wood Technol., nr 69.

14.

EN 438-2:2005 High- pressure decorative laminates (HPL) Sheets based on thermosetting resins (usually called laminates).

15.

Kozakiewicz, K. (1999) Badania właściwości sorpcyjnych i powierzchni wewnętrznej drewna z sarkofagów egipskich, praca magisterska, Wydział Technologii Drewna SGGW [Tests of Sorption Properties and Internal Surface of Egyptian Sarcophagi Wood, M.S. thesis, Faculty of Wood Technology, Warsaw University of Life Sciences], Warszawa.

16.

Ławniczak, M. (1972) Zarys hydrotermicznej i plastycznej obróbki drewna, Skrypt AR w Poznaniu [Overview of Hydrothermal and Plastic Wood Processing, Script of Poznań Academy of Life Science].

17.

Matejak, M. (1983) Primare Sorptionsisothermen von Holz, Holzforschung und Holzverwertung 35 (Februar): 1-6.

18.

Nikitin, N.J. (1955) Chemia drewna i celulozy [Wood and Celulose Chemistry], Warszawa.

19.

Kübler, H (1962) Schwinden und Quellen des Holzes durch Kalte, Holz als Roh- und Werkstoff 20: 364-368.

20.

Cudinow, B.S. (1967) Uber den Warmeaustausch im Holz beim Gefrieren, Holztechnol 8: 93-96.

21.

Wanin, S.I. (1953) Nauka o drewnie [Wood Science], Warszawa

22.

Matejak, M., Starecka, D. (1971) Einfluss des Gerfieren von Holz auf seine Druckfestigkeit, Holztechnologie 3: 144-146.

23.

Kollmann, F. (1951) Technologie des Holzes und der Holzwerkstoffe, München

24.

Kollmann, F., Engel, D. (1965) Anderungen der chemischen Zusammensetzung von Holz durch thermische Behandlung, Holz als Roh- und Werkstoff 23, 12: 461-468.

25.

Jarmutowska, A. (1973) Zmiany w składzie i strukturze węglowodanów zawartych w drewnie świerka zachodzących w określonych warunkach sztucznego starzenia, rozprawa doktorska, [Changes in Spruce Wood Carbohydrate Composition and Structure Taking Place in Certain Artificial Ageing Conditions, Doctoral Thesis], SGGW-AR, Warszawa

26.

Krzysik, F. (1978) Nauka o drewnie [Wood Science], Warszawa

27.

Helińska-Raczkowska, L., Raczkowski, J. (1971) Niektóre zagadnienie przyspieszonego starzenia drewna, Rocznik WSR w Poznaniu 6. [Some Issues Concerning Accelerated Wood Ageing]

10

217


218

7. The Sawing Process Oral Presentations


Economical Wood Sawing With Circular Saw Blades Wasielewski, Roman1, Orlowski, Kazimierz1( ) and Szyszkowski, Stanislaw2 1

Gdansk University of Technology, Faculty of Mechanical Engineering, Department of Manufacturing Engineering and Automation, Gdansk, Poland 2 PP GASSTECH Sp. z o. o. (PLC), Suwalki, Poland 1 [email protected], 2 [email protected] ( ), 3 [email protected]

ABSTRACT Material- and energy-saving belong to the basic requirements which are imposed on the contemporary manufacturing processes. Since realization of those processes gives measurable profits not only economical but also ecological. In case of wood sawing with circular saw blades material- and energy-savings are dependent on total overall set of teeth (theoretical kerf) and teeth position accuracy in relation to the workpiece. Hence, obtaining a decrease of both raw material and energy losses the use of narrow-kerf saw blades, an increase of sawing accuracy, and besides a reduction of spacing, in the case of sawing with a gang of circular saws, are required. However, an intention materialization of those requirements in the case of sawing with circular saw blades in a quite difficult issue and depends on many factors. The carried out detailed analyses concerning ways of chip transportation in the kerf slot, saw blade stiffness, saw blade movement and workpiece feeding accuracy are the proofs of those inconvenient technical problems for sawmills. Not till then the whole system of sawing fulfils defined requirements the realization of economical wood sawing with circular saw blades seems to be practical.

INTRODUCTION The ecological friendship is a basic demand, which is placed for contemporary manufacturing processes. In the field of wood sawing with circular saw blades, which belongs to basic cutting methods in the wood industry, the ecological friendship first of all is connected with material and energy-saving of the process. The application of the technologies in sawmills, which allow the users to reduce raw material losses and energy consumption, gives measurable advantages not only economical but also ecological [1, 2]. Environmental restrictions and increasing log costs have caused many sawmills to look at new ways to extract more value from their raw material. One of the more traditional ways to accomplish this aim is to increase volume recovery. Firstly, a saw kerf can be reduced through improvements in saw design that reduce either the saw blade thickness or the side set of the saw [3, 4]. However, it has been shown that changes in these two saw design factors can lead to increased within-board sawing variation, or deviation through the cut [3]. Those phenomena may be caused by decreasing either the saw blade specific stiffness or the saw blade operating stiffness due to the loss of saw blade stability [1]. The latter also may be an effect of the temperature increase owing to passage of chips between the saw blade and kerf walls [5]. Twin shaft multi-rip saws make possible very efficient sawing of wood even in the case of large cutting depths. With regard for that those circular sawing machines are willingly and often used in sawmills of large productivity. On those machine tools collar clamped circular saw blades of the “Multix” type are frequently applied. They are designated for ripping of hard and soft green wood. Except for traditional carbide teeth on the rim, the saw is equipped additionally with four scraper carbide edges (cleaning knives) (Figure 1a). Producers of those tools advertise that the

219


“Multix” circular saws provide fast removal of shavings from cut space. Nevertheless, there are known some cases of catastrophic damages of saw blades due to the passage of chips between the saw blade and kerf walls [5]. For this reason a new circular saw design has been developed (Figure 1b). The new circular saw has larger static and dynamic stiffness, hence, protects a saw blade against deviations through the cut. Furthermore, special scrapers, situated just below each tooth, prevent the saw blade from heating by the uncontrolled chip flow [4, 6]. The goal of this paper is to make the comparative assessment of both the traditional collar clamped circular saw blades of the “Multix” type and the new design of the circular saw “Ekomultiks” in industrial plant conditions. b)

a)

Figure 1: Circular saw blades for rip sawing of the traditional issue of a type “Multix” (a) and of a new design “Ekomultiks” (b)

RELATIVE KERF LOSSES AND RAW MATERIAL YIELD The individual sawmill decides where its strong points are and how best to choose the most economical way to improve productivity. Where one sawmill may prefer to improve its efficiency, another might concentrate its efforts on better utilisation of raw materials. Operational reliability, however, is by far the most important factor [7]. At the preliminary stage of the assessment of the effect of the kerf size upon raw material savings may be done by calculation of relative kerf losses Qm, which are estimated in the case of sawing machines, for a gang of saws, as a ratio of the total loss volume and the stock volume [1, 8], on the assumption that each board has the same thickness: j =2

(n + 1) R + ∑ Gbj Qm =

j =1

(1)

j =2

(n + 1) R + ∑ Gbj + n ⋅ Gmin j =1

where: n – number of produced boards, R – maximum value of the real kerf calculated as R = S t + B , where St is a sum of theoretical kerf (overall set), B is an axial run out in relation to the workpiece (equal to roughness of sawn surface Rt, figure 2), Gmin is a actual minimal board thickness value, and Gbj is thickness of side boards (slabs or offcuts). If in the sawn pattern

220


elements have different thickness instead of n ⋅ Gmin put

j =n

∑G j =1

min j

into Eq. (1). In Figure 2 are

shown two cases of real kerfs observed on board surface, the kerf #1 while changes of the sawn profile depend on circular saw blade rotational frequency, and the kerf #2 when changes of the circular saw blade position are slow (snaking, wandering of the circular saw blade). In some sawmills determination of raw material yield is more frequently in use because it illustrates which part of the raw material is utilised as ready-made products. The raw material yield Wm for the analysed sawing pattern can be determined as follows: j =2

(n + 1) R + ∑ Gbj Wm = 1 − Qm = 1 −

j =1

(2)

j =2

(n + 1) R + ∑ Gbj + n ⋅ Gmin j =1

R

G* min G* max

St

B/2

Gmin

Rt2 Rt1

Gmax

2

1

Figure 2: Kerf and board thickness changes, where G*min and G*max are acceptable board thickness values, Gmin, Gmax are minimal and maximal values of actual thickness In case of raw material which has equal both width G and height H, on its whole length L, the assessment task seems to be very easy. However, in industrial reality ready-made elements are often obtained from raw material which width and height change along its length (Figure 3). The dimension D is the smallest width of the plank with height H on both sides (Figure 3a). In the batch of raw material which comes to the rip saw, the smallest dimension D can change in a range of . During cutting of a large batch it can be assumed that probability of rip sawing of planks with size D is in the range of and has a linear distribution (Figure 3b). In case of determined planks when boards with total width of Gc should be obtained only W% of elements A can achieve height H along the whole plank length L. The use of the sawing technology in which the total width of Gc’ is needed, the quantity of elements A with dimension H along the whole plank length L will increase in value of ∆W % (Figure 3d). That increase of element A efficiency, for Dmin < Gc < Dmax can be determined as:

∆W =

Gc − Gc' ∆Gc ⋅ 100 = ⋅ 100% Dmax − Dmin Dmax − Dmin

221

(3)


During rip sawing operations of the plank in shape as shown in Figure 3 the increment of raw material yield ∆W depends on both the necessary total width Gc of sawn boards and on the spread of the dimension D range in the batch.

a)

Gc D

percentage distribution of dimension D

100 [%]

H

b)

50

A

L

W

c)

W

D [mm]

0 Dmin Dmax

Gc Gc'

Gc/2

d)

D

Figure 3: Changes of width and height of side offcuts along the plank length (a), percentage change distribution of the smallest width D (b), determination of material yield for defined total width Gc (c) and the effect of the total width reduction on the material yield (d) Thus, a reduction of material losses of the sawing process with circular saw blades demands for: a reduction of the tooth overall set of the circular saw (application of thinner saw blades), an increase of sawing accuracy (reduction of circular saw blades axial run-out) and a reduction of the saw blade spacing. The effects of those actions are presented in Figure 4, in which an evident proof of application of a new design of circular saws is visible. It is especially perceptible if lamellae are an effect of rip-sawing. In the case presented in Figure 4 the raw material yield Wm increased about 10%, there was about 16% less chips, hence, there was more wood from side boards which could be raw material for production of chips for instance. Furthermore, it could be expected a reduction of power consumption about 16% [9].

MATERIAL AND METHODS Circular saws of the new design (“Ekomultiks”, Figure 1b) and the traditional issue (“Multix” type, figure 1a), both types with carbide tipped teeth, were examined in industrial plant conditions on the twin shaft multi-rip saw Heavy Duty PRW422 (f. TOS Svitavy, CZ). Circular saw blades’ technical data: -

the traditional issue “Multix”: diameter Ø300 mm, overall set (kerf) St = 3.7 mm, saw blade thickness s = 2.5 mm separated with distance pieces of Ø110 mm (nominal controlled board thickness 25.3 mm);

222


-

the new design “Ekomultiks”: diameter Ø300 mm, overall set (kerf) St = 3.4 mm, saw blade thickness s = 2.5 mm separated with distance pieces of Ø110 mm (nominal thickness of 25.6 mm for controlled boards)

-

the new design “Ekomultiks”: diameter Ø300 mm, overall set (kerf) St = 3.1 mm, saw blade thickness s = 2.2 mm separated with distance pieces of Ø130 mm (nominal thickness of 25.1 mm for controlled boards).

11 pcs.

10

Figure 4: Industrial examples of sawing of similar timber for the same dimensions of the obtained sawn lamellae, where: upper timber was slotted with the use of circular saw blade “Ekomultiks” whereas lower was sawn with circular saw blade “Multix” applied [9] Planks (Pinus sylvestris L.) of moisture content MC of 25–32% were sawn. During one 8 hour shift, sawn lumber at the controlled position in the gang (board #2, figure 4), was measured with the digital caliper (f. Gedore). Measurements were executed on the upper (Gg) and lower measurement line (Gd) respectively.

RESULTS AND DISCUSSION Comparison of pine planks slotted during carried out industrial experiments with gangs of circular saw blades in case of sawing with traditional circular saws “Multix” (lower plank) and new circular saws “Ekomultiks” (upper plank) is presented in Figure 5. In this picture the position of the measured board is shown as #2. In Figure 6 results of board thickness measurement determined on the lower and the upper measurement line for rip-sawing with “Multix” circular saw blades and “Ekomultiks” circular saw blades are shown. While the “Multix” circular saw blades are applied there was observed the range of lumber thickness distribution equal to 0.5 mm. In the next step “Multix” circular saw blades were replaced with “Ekomultiks” circular saw blades (saw blade thickness of 2.5 mm) separated with old distance pieces. In that conditions the lumber thickness distribution equal to 0.3 mm both on the upper and lower line was obtained. Eventually, we have decided to reduce both circular saw blade thickness (s = 2.2 mm) and distance pieces’ thickness, and a result of

223


effects of sawing with “Ekomultiks” circular saw blades separated with new distance pieces (Ø130 mm) are performed in Figure 6c. It should be emphasized that a reduction of saw blade thickness and simultaneous an increase of the distance piece diameter (from Ø110 mm to Ø130 mm) has guaranteed the same value of the circular saw blade stiffness.

#2 New circular saws

Old circular saws

#2

Figure 5: Comparison of pine planks slotted with gangs of circular saws in case of sawing with traditional circular saws “Multix” (lower plank) and new circular saws “Ekomultiks” (upper plank); where: #2 are controlled boards

Figure 6: Board thickness determined on the lower and the upper measurement line for sawing with “Multix” circular saw blades (St = 3.7 mm) (a), “Ekomultiks” circular saw blades (St = 3.4 mm)with the use of old distance pieces (b) and “Ekomultiks” circular saw blades (St = 3.1 mm)with the use of a new type of distance pieces (c), where: GdN, GgN – board nominal thickness, Gd , Gg – board average thickness

224


Figure 7: Distribution of changes of the input workpiece width and circular saw blade spacing for sawing with “Multix” circular saw blades (A), and “Ekomultiks” circular saw blades (B) Additionally comparison of effects of the material-saving to the traditional sawing technology may be done analytically on the basis of distribution of the input workpiece width changes (Figure 7). In those sawing conditions, from one hundred sawn pine planks have been obtained: 2×(76.9 + 35.5) = 224.8 pieces of side lumber in the case A, and in the case B 2×(82.4 + 44.2 + 6.1) = 265.4 items of the side lumber. Thanks to application of the pro-ecological technology there was also achieved: about 18% increase in the amount of the side lumber (according to Eq. (3)), roughly 16% less sawdust (as an effect of kerf reduction) and about 16% a lower values of the cutting power consumption.

SUMMARY The use of narrow-kerf saw blades and an increase of the sawing accuracy reduces the both cutting losses and cutting energy consumption while sawing of wood. The better use of the input raw material in the presence of the lower cutting energy consumption is the basis of the proecological technologies of wood sawing. The application of such a kind technology allows the sawmillers to reduce raw material consumption and furthermore brings measurable economical profits.

ACKNOWLEDGEMENT The authors1,2 would like to thank the firm PPH GASSTECH Ltd. (Suwalki, PL) for the donation of the circular saw blade used in tests. They also would like to acknowledge firms PPH GASSTECH Sp. z o.o. (PLC, Suwalki, PL) and Wydawnictwo Inwestor Sp. z o.o. (publisher, PLC, Tczew, PL) for their financial support for the author2 participation at the 20th International Wood Machining Seminar. Furthermore, we also have to thank the firm P.U.P "COMPLEX" Sp. z o.o (PLC, PL) which has allowed us to conduct experiments at its production sawmills in Dziemiany and Trzebuń. The authors would like to acknowledge that the circular saw blade

225


“Ekomultics” was awarded with the Gold Medal at the International Trade Fair of Machines and Tools for the Wood and Furniture Industries “Drema 2010” in Poznan (PL).

REFERENCES 1. ORLOWSKI, K. (2003) Materiałooszczędne i dokładne przecinanie drewna piłami. (In Polish: Narrow-kerf and accurate sawing of wood). Seria Monografie nr 40, Wydawnictwo Politechniki Gdańskiej, Gdańsk. 2. ORLOWSKI, K., WASIELEWSKI, R., SZYSZKOWSKI, S., WNUKOWSKI, E. (2007) The effect of improved cutting conditions of the circular saw blade on precision of cutting. Ann. WULS-SGGW, For and Wood Technol. No 62: 100–104 3. MANESS, T.C., LIN, Y., (1995) The Influence of Sawkerf and Target Size Reductions on Sawmill Revenue and Volume Recovery. Forest Prod. J., 45(11/12): 43–50. 4. WASIELEWSKI, R., ORŁOWSKI, K., SZYSZKOWSKI, S., WNUKOWSKI, E. (2007) ZGŁOSZENIE PATENTOWE (Patent pending) P.382274, Piła tarczowa z rowkami wiórowymi. (In Polish: Circular saw blades with chip spaces) (2007.04.24). 5. WASIELEWSKI, R. (2009) Influence of chip transport method on effects of cutting with circular saw. pp. 44–59. In: Górski, J., Zbieć, M. (Eds.) Wood machining and processing – product quality and waste characteristics. WULS–SGGW, Warsaw. 6. WASIELEWSKI, R., ORŁOWSKI, K., SZYSZKOWSKI, S., WNUKOWSKI, E. (2008) Piły tarczowe o podwyższonej sztywności. (In Polish: Circular saw blades with stiffness increased) pp. 426–433. In: Stós, J. (Ed.) Obróbka skrawaniem – Innowacje, Szkoła Obróbki Skrawaniem ; 2. Kraków : IOS Instytut Zaawansowanych Technologii Wytwarzania. 7. SANDVIK (1999) Production, use and maintenance of wood bandsaw blades. A manual from Sandvik Steel. AB Sandvik Steel, Sandviken, Sweden, May, S–336–ENG. 8. WASIELEWSKI, R. (2010) Losses and raw material yield of wood sawing processes. Ann. WULS-SGGW, For and Wood Technol. 72. No 72: 414–417. 9. WASIELEWSK, R., ORŁOWSKI K. (2010) Pro-ecological technology of wood sawing with circular saw blades. Ann. WULS-SGGW, For and Wood Technol. 72: 423–426.

226


Improvement of sawing efficiency in sawing frozen wood Ikami Yuji, Murata Kohji Forestry and Forest Products Research Institute, Tsukuba, JAPAN

ABSTRACT Frozen wood decreasing sawing efficiency at sawmills in the northern part of Japan has become a problem. To date, the main focus has been on improved band saws and/or changed sawing conditions as potential remedial measures when sawing frozen wood. Conversely, the number of sawmills introducing a boiler for dry-kilns and/or a power plant has been increasing in recent years and the use of surplus heat generated from such facilities might reveal a new solution to the problem. In this study, we conducted a sawing test comparing the sawing forces required for frozen and defrosted wood to determine fundamental data showing the effect of the method of applying surplus heat to defrost the frozen wood on sawing efficiency improvement when sawing frozen wood. When sawing work pieces obtained from sugi (Cryptomeria japonica D. Don) logs including low-moisture heartwood, the sawing force required for sapwood increased greatly when frozen, whereas the increase for heartwood was slight. After defrosting for 4 hours, the sawing force required for sapwood decreased drastically, and after 6 hours, it declined even further until approaching the value for unfrozen work pieces. When sawing the work pieces obtained from sugi logs including high-moisture heartwood, the sawing force required for both sapwood and heartwood increased significantly due to freezing. The sawing force required for sapwood decreased drastically after defrosting for 4 hours, and approached the value for unfrozen work pieces when defrosted for 6 hours. Although the sawing force required for heartwood decreased gradually, it remained above the value for unfrozen work pieces, even after defrosting for 6 hours.

INTRODUCTION Sawing frozen wood in the northern part of Japan has proved problematic since it results in sawing inaccuracy, decreased sawing efficiency due to slower feed speed, and increased power consumption, etc. To date, the main focus has been on improved band saws and/or changed sawing conditions (feed speed or saw wheel rotation speed) as remedial measures when sawing frozen wood. In fact, some sawmills use special band saws, for instance, the perforated or with a small kerf width, to limit any decline in sawing efficiency as far as possible. Conversely, the number of sawmills introducing a boiler for dry-kilns and/or a power plant has been increasing in recent years and the use of surplus heat generated from such facilities might reveal a new solution to the problem. With this in mind, we started a series of research to improve sawing efficiency after defrosting the frozen wood with said surplus heat. A cooperating sawmill in Gifu Prefecture, where a gasification CHP (Combined Heat & Power) plant was installed, was selected. Although the sawmill converts about 20,000m3 sugi (Cryptomeria japonica D. Don)

227


logs into sawn lumber each year, the productivity in winter declines due to the sawing of frozen wood. Figure 1 shows the power consumption when sawing frozen and unfrozen wood using the twin band mill at the sawmill. The feed speeds when sawing frozen and unfrozen wood were 9.4 and 32.6 m/min, respectively, while the average integrated power consumption per single sawing process when sawing frozen and unfrozen wood were 0.03 and 0.008 kWh, respectively. Because of the drastic decline in sawing efficiency, the sawmill set up a simple log stock house to avoid the logs becoming snow-covered in winter. In view of the sawmill circumstances, we considered the potential for warming the logs stocked in the house via the surplus heat of the gasification plant. Initially, we attempted to circulate the hot water generated from the gasification plant via a polyethylene pipe installed in the log stock house and could already confirm the potential to increase the temperature in the log stock house (see Figure 2). As the next step in this study, we conducted a sawing test comparing the sawing forces required for frozen and defrosted wood to determine fundamental data showing the effect of the method of applying surplus heat to defrost frozen wood on sawing efficiency improvement when sawing frozen wood. ②

②

①

③

①

135 cm ×165 cm

③

Powe r con sumption (kW )

Sawing pattern 40 Frozen wood 30 20

②

①

10

③

0

Power con sump tion (k W )

0

20

40

60 Time (sec)

80

100

40 Unfrozen wood 30

②

20

①

③

10 0

0

10

20 Time (sec)

30

40

Figure 1. Power consumption when sawing frozen and unfrozen wood.

228


Polyethylene pype Hot water

Log stock house

Gasification CHP plant

Figure 2. Consept of the utilization of surplus heat.

Experiment Work pieces 100 mm thick, 200 mm long, and 200 mm wide were obtained from two types of sugi log, namely, logs including low-moisture and high-moisture heartwood (Types A and B). For the work pieces obtained from Type A logs, the average moisture contents and specific gravities, oven-dry, for sapwood and heartwood were 218.5％ and 0.28, 66.7％ and 0.31, respectively. For those obtained from Type B logs, the equivalent figures were 220.2％ and 0.30, 209.0％ and 0.34, respectively. The frozen work pieces were prepared by freezing in a freezer for more than 24 hours at -12 ℃. Two kinds of defrosted work pieces were prepared after defrosting the frozen work pieces in an incubator for 4 and 6 hours respectively at 20 ℃. The work pieces were sawn into 5 mm thick slats at three different feed speeds of 20.0, 30.0, and 40.0 m/min with the constant saw wheel rotation speed 700 rpm, using a 1,100 mm band mill with a light auto-feed carriage. The specifications of the band saw used, namely, blade width, blade thickness, tooth pitch, clearance angle, sharpness angle, and rake angle were 127 mm, 1.05 mm, 32 mm, 22°, 43°, and 25°, respectively. The parallel sawing force at the sapwood and heartwood portions was detected with a cantilever type load cell installed on the carriage (see Figure 3). Sapwood

Heartwood

F ee d

Parallel sawing force

Figure 3. Load cell for measuring sawing force.

229


RESULTS AND DISCUSSION Figure 4 shows the variation in parallel sawing force for the work piece obtained from Type A logs with the degree of freezing. When sawing unfrozen work pieces, the sawing force required for heartwood exceeded that of sapwood. This is influenced by the fact that the moisture content of heartwood is less than that of sapwood and the specific gravity of heartwood is larger than that of sapwood. The sawing force required for sapwood increased greatly by freezing, whereas the increase in that required for heartwood was slight. Sawdust adhering to the sawn face was observed when sawing the sapwood portion of the frozen work piece (see Figure 5). In previous Sapwood

Sa wing fo rce (N)

120

Heartwood Feed speed : 20m/min

100 80 60 40 20 0 140

Feed speed : 30m/min

Sawing force (N)

120 100 80 60 40 20 0 160


Sa wing fo rce (N)

140 120 100 80 60 40 20 0

Unfrozen

Frozen

Defrosted Defrosted for 4 hours for 6 hours

Figure 4. variation in parallel sawing force for the work piece obtained from logs with low-moisture heartwood with the degree of freezing.

230


Figure 5. Sawdust adhering to the sawn face.

studies, the power consumption when sawing frozen wood with high-moisture was reportedly larger than that for low-moisture［1,2］. Since much of the frozen wood with high-moisture is frozen, the cutting resistance is increased. The considerable sawing force required for sapwood was thought to be affected by its high moisture content. After defrosting for 4 hours, the sawing force required for sapwood decreased drastically, and after 6 hours, it declined even further until approaching the value for unfrozen work pieces. A similar tendency was seen at all feed speeds. Figure 6 shows the variation in parallel sawing force for the work piece obtained from Type B logs with the degree of freezing. The sawing force values for the Type B work pieces overall exceeded those for Type A work pieces. Both the sawing force required for sapwood and that for heartwood increased greatly by freezing, with the increased sawing force required for heartwood especially prominent. Sawdust adhering to the sawn face was observed when sawing both the sapwood and heartwood portions of frozen work pieces. After defrosting for 4 hours, the sawing force required for sapwood decreased drastically, and after 6 hours, it declined even further until approaching the value for unfrozen work pieces. Although the sawing force required for heartwood decreased gradually, it remained above the value for unfrozen work pieces, even after defrosting for 6 hours. A similar tendency was seen at all feed speeds. It is thought that a longer duration or higher temperature defrosting procedure is needed, because this phenomenon showed that the central portion of the work piece had not been sufficiently defrosted. Based on these results, the method of applying surplus heat for defrosting the frozen wood was considered effective in terms of improving efficiency when sawing frozen wood. For example, it is expected that afternoon sawing efficiency will be improved by sawing logs put in the log stock house in the morning. We are going to conduct a sawing test measuring the other parameters (power consumption, roughness of sawn face, etc.) or using larger work pieces to further improve this method and its practicality.

231


Sapwood

Heartwood

120 Feed speed : 20m/min

Sa wing fo rce (N)

100 80 60 40 20 0

Sawing force (N)

200


160 120 80 40 0

Sa wing fo rce (N)

300


250 200 150 100 50 0

Unfrozen

Frozen

Defrosted Defrosted for 4 hours for 6 hours

Figure 6. variation in parallel sawing force for the work piece obtained from logs with high-moisture heartwood with the degree of freezing.

232


CONCLUSION We conducted a sawing test comparing the sawing force required for frozen and defrosted wood. When sawing the work pieces obtained from logs with low-moisture heartwood, the sawing force required for sapwood increased greatly when frozen, whereas the increase for heartwood was slight. After defrosting for 4 hours, the sawing force required for sapwood decreased drastically, and after 6 hours, it declined even further until approaching the value for unfrozen work pieces. When sawing the work pieces obtained from logs with high-moisture heartwood, both the sawing force required for sapwood and that for heartwood increased greatly by freezing. After defrosting for 4 hours, the sawing force required for sapwood decreased drastically, and after 6 hours, it declined even further until approaching the value for unfrozen work pieces. Although the sawing force required for heartwood decreased gradually, it remained above the value for unfrozen work pieces, even after defrosting for 6 hours. It was thought that the method of applying surplus heat to defrost the frozen wood may be effective in improving sawing efficiency when sawing the frozen wood.

REFERENCES 1. Kitazawa N, Yanagizawa Y, Takasu N (1968) Sawing of the Frozen Log (Ⅰ). Journal of the Hokkaido Forest Products Research Institute, 9-12 2. Maeda I, Takasu N, Yanagizawa Y (1971) Sawing of the Frozen Log (Ⅱ). Journal of the Hokkaido Forest Products Research Institute, 6-9

233


Influence of Sawing Patterns on Lumber Quality in Large Sugi Logs Matsumura Yukari, Murata Kohji, Ikami Yuji Forestry and Forest Products Research Institute, Matsunosato 1, Tsukuba, Ibaraki, JAPAN

ABSTRACT Appropriate log sorting and optimal sawing patterns are indispensable to the efficient production of lumber. Our previous studies showed the potential to produce lumber suitable for end use efficiently by sorting logs considering Young's modulus and moisture content and sawing them using optimal sawing patterns. It is important to clarify the relationship between the quality of logs and that of sawn lumber to determine an effective sorting and optimal sawing pattern. Although some information exists about lumber quality in medium sugi (Cryptomeria japonica D. Don) logs, little is known about lumber quality in large logs, the supply of which is expected to increase imminently. Therefore, in this study, we investigated the influence of the sawing pattern of large sugi logs on the relationship between the log and lumber quality. Large sugi logs (30-40 cm diameter) grown in Tochigi Prefecture were converted into sawn lumber using different sawing patterns, the main products of which were squared lumber with pith, squared lumber without pith, and scantling. We focused on the relationship between Young’s modulus of log and that of lumber, as well as that between log sweep and lumber warp. The sawing yields varied for each sawing pattern. There is the potential to produce reliable lumber by considering the variation of Efr within a log and the difference of lumber warp according to sawing patterns.

INTRODUCTION Japan had a forest area of 25 million hectares in 2007, which comprised about 67% of the country. The growing stock is 4.4 billion m3, which increases by 80 million m3 annually, and about 10 million hectares of forest area is softwood forest planted to produce lumber for wooden housing. In 2006, the ratio of more than ten age classes in planted forest area was about 35%, and will increase to about 60% in the coming decade [1]. As the ratio of the forest area of high age classes increases, the supply of large logs is also expected to expand and it is important to ensure such mature forest resource is used sustainably. The majority of both the planted area and stock of forest in Japan are sugi (Cryptomeria japonica D. Don), which must therefore be used effectively. Appropriate log sorting and optimal sawing patterns are indispensable to the efficient production

234


of lumber. Our previous studies showed the potential to produce lumber suitable for end use efficiently by sorting logs considering Young’s modulus and moisture content and sawing them using optimal sawing patterns [2] [3]. Although some information exists about lumber quality in medium sugi logs, little is known about that in large logs, the supply of which is expected to increase imminently. The variation of the sawing pattern increases, meanwhile, with increasing log diameter. The purpose of this study is to examine the influence of sawing patterns on the lumber quality of large sugi logs. In this study, we investigated the relationship between the log and lumber quality for several sawing patterns, while the volume and value yields for each were also calculated.

MATERIALS AND METHODS Large sugi logs grown in Tochigi Prefecture were used in this study. The total was 141, ranging in diameter from 30 to 40 cm and 3.65 m long. After debarking, the lengths, diameters of the top and butt ends, number of annual rings, and the percentage of heartwood were all measured. The logs were graded based on the Japanese Agricultural Standard (JAS) [4]. The natural frequency of vibration of log was measured by a longitudinal vibration method and Young’s modulus of the logs was calculated by Equation (1), E fr  4  L2  fr 2  

(1)

where, Efr (GPa) is Young's modulus of the logs, L (m) the length, fr (Hz) the natural frequency of vibration, and ρ (kg/m3) the bulk density. A 1,200 mm band mill with an auto feed carriage and a 1,100 mm auto-roller table band resaw were used in this study. Logs were sawn using the four sawing patterns shown in Figure 1. The main products of sawing pattern 1 were flat square lumber, 25.0 cm wide and 13.0 cm thick. The main products of sawing pattern 2 were squared lumber with pith, 11.5 cm wide and 11.5 cm thick, while those of sawing pattern 3 were scantling 11.5 cm wide and 3.8 cm thick and for sawing pattern 4, squared lumber without pith, 13.0 cm wide and thick. In sawing pattern 4, several pieces of squared lumber were sawn according to the diameter of the logs. The sawn lumber was graded based on the Japanese Agricultural Standard (JAS) [5] immediately after sawing. The main products of each sawing pattern were weighed and had their width, thickness, and length measured. The warp of the lumber was measured in units of 1 mm using a thread and ruler, while the Young’s modulus of the main products was calculated using the longitudinal vibration method. The volume and value yields in each sawing pattern were calculated by Equations (2) and (3). Yvol 

 V   100 i

V0

(2)

 V Yval    i  V0

  Pi      P0

    100  

(3)

where, Yvol (%) is the volume yield, Yval (%) the value yield, Vi (m3) the volume of each sawn lumber, Vo (m3) the volume of the log, Pi (yen/m3) the price of each sawn lumber, and Po (yen/m3) the standard lumber price.

235


Fig. 1 Sawing patterns used.

RESULTS AND DISCUSSION Specification of the logs used Table 1 shows the specification of logs used for each sawing pattern. In all logs used, the average diameter of the top end was 33.0 cm, that of the butt end was 38.5 cm, the volume was 0.414 m3, the number of annual rings was 52, and the percentage of heartwood was 47.7%. Table 1 Specifications of the logs used. Diameter

Sawing Number pattern

of logs

1

53

2

26

3

37

4

25

Top end Butt end (cm)

(cm)

Length (cm)

Volume 3

Weight

(m )

(kg)

Annual

Heartwood

rings

(%)

Young's modulus (GPa)

AVE.

32.5

37.4

376.3

0.400

265.6

58

48.0

8.16

SD

2.4

3.6

4.0

0.063

53.4

17

8.4

0.81

AVE.

33.8

40.6

374.1

0.429

283.2

54

51.9

6.14

SD

2.3

3.9

2.2

0.058

56.0

13

8.9

0.63

AVE.

33.0

39.1

381.3

0.416

256.5

42

44.9

6.07

SD

2.0

2.8

3.5

0.053

42.6

3

7.7

0.51

AVE.

34.1

38.4

374.2

0.422

264.7

50

46.9

7.28

SD

2.4

3.2

1.4

0.066

47.0

5

6.0

1.01

236


Young’s modulus of log and lumber Figure 2 shows the relationship between Young’s modulus (Efr) of the logs and that of the main products for each sawing pattern. The coefficients of determination were high at sawing patterns 1, 2, and 4, which indicate that sorting logs by Efr would be useful for the efficient production of reliable sawn lumber from large logs similarly to those of medium logs. The strength of the correlation differed according to sawing pattern. In this experiment, of the three sawing patterns, the sawing pattern 4 had the strongest correlation between the Efr of the log and the squared lumber without pith, of which the former shows the average Young's modulus of the whole log. Therefore, if the dimensions of the main products were large, the Efr of lumber would indicate a similar Efr of logs.

Fig. 2 Relationship between the log and lumber Efr.

Conversely, the coefficients of determination were low at sawing patterns 3. In general, Efr varies within the log according to its position in the same. This variation was reflected in the difference of each Efr of lumber, due to the small dimensions of the lumber in sawing pattern 3. Consequently, there was little correlation between the Efr of log, showing the average Efr for the whole log and the Efr of lumber, which varies widely in sawing pattern 3. We examined the influence of the lumber position on the log on the variation of Efr of lumber. The positions of the lumber on the log in sawing pattern 3 were illustrated in Figure 3. Here, “center” is the position including the pith or a part thereof; “outer” is the outermost position of the log and “inner” is a position between the two. The respective totals for lumber pieces corresponding to “center”, “inner”, and “outer” were, 47, 148, and 70. The frequency distributions of the lumber Efr for each position are represented in Figure 3. The average for

237


lumber Efr corresponding to “center”, “inner”, and “outer” were respectively 4.64, 5.43, and 6.48 GPa, with the average Efr differing significantly at the 1% level. This tendency resembles the result of other studies, whereby that specific MOE increased significantly from pith to bark in sugi [6]. It is important to note the variation of Efr in logs when several main products were sawn as in sawing pattern 3, especially when requiring lumber strength.

Fig.3 Comparison of thefrequency distribution of lumber Efr in sawing pattern 3 for each position on the log.

Lumber warp Figure 4 shows the warp degrees for flat square, squared lumber with pith, and squared lumber without pith. Here, Wa and Wb are the parallel and vertical warps with log sweep. The previous investigation showed that the warp degree of lumber pieces sawn from logs with greater sweep tends to be larger than for lumber sawn from logs with less sweep [7]. However in this study, no such tendency emerged between the log sweep and the warp degree of lumber. It is reasonable to consider the influence of log sweep small, due to the large diameter and comparatively small sweep of the logs. Conversely, there were significant differences among the average Wa of the flat square, squared lumber with pith, and squared lumber without pith at a 1% level probability, as well as among the average Wb. Of the three products, the squared lumber without pith had the largest warp degree, and the flat square the lowest. The warp of the lumber is likely to be influenced by the sawing patterns. In addition, Wa and Wb of squared lumber without pith differed significantly at a 1% level probability. These facts suggest that the lumber warp was affected by the sawing pattern rather than the log sweep. There is a possibility of reducing the lumber warp by using a sawing pattern taking the variation of lumber warp caused by position on the log into consideration.

238


Fig. 4 Warp degree of lumber in sawing patterns 1, 2, and 4.

Sawing yields Figure 5 shows the average volume and value yields for each sawing pattern, which differed significantly at a 1% level of probability. In general, the volume yield tends to be high in the sawing pattern to produce large dimension lumber, because the decrease caused by kerf is small. Additionally, the value yield for logs with a high volume yield tends to also be high when the log grade is at the same level. Of the four sawing patterns, sawing pattern 1 producing a flat square, had the highest volume yield, while sawing pattern 3, producing scantlings, had the lowest. The value yield showed a similar tendency to volume yield.

Fig. 5 The average volume and value yields for each sawing pattern.

239


CONCLUSION This study was intended to examine the influence of sawing patterns on the lumber quality of large sugi logs and in it, we investigated the relationship between logs and lumber quality in several sawing patterns. The result of the experiment shows that sorting logs by Efr would be useful for the efficient production of reliable sawn lumber from large logs, with large dimensions of main products, such as sawing patterns 1, 2, and 4. Conversely, it is important to note the variation of Efr in logs when several main products were sawn as sawing pattern 3, especially when lumber strength is required. The lumber warp was affected by the sawing pattern rather than the log sweep, but may be reduced by using a sawing pattern considering the variation of lumber warp caused by position on the log. The sawing yields varied for each sawing pattern. There is the potential to produce reliable lumber by considering the variation of Efr within a log and the difference of lumber warp according to sawing patterns.

REFERENCES 1. Ministry of Agriculture, Forestry and Fisheries, Japan (2009) Annual Report on Trends in Forest and Forestry in Japan. 2. Matsumura Y., Murata K., Ikami Y. (2005) Influence of Sorting Logs By Young's Modulus and Moisture Content on Sawn Lumber Yields and Qualities. Proceedings of 17th International Wood Machining Seminar, 2:481-488 3. Matsumura Y., Murata K., Ikami Y. (2007) Effects of sorting logs on sawn lumber yields and qualities -Young's modulus and moisture content-. Bulletin of FFPRI, 6(1):1-7 4. Ministry of Agriculture, Forestry and Fisheries, Japan (2007) Japanese Agricultural Standards for Log. 5. Ministry of Agriculture, Forestry and Fisheries, Japan (2007) Japanese Agricultural Standards for Sawn lumber. 6. Zhu J., Takata K., Iijima Y., Hirakawa Y. (2003) Grouth and Wood Quality of Sugi (Cryptomeria japonica D. Don) Planted in Akita Prefecture 1. Mokuzai Gakkaishi, 49(2):138-145 7. Ikami Y., Matsumura Y., Murata K., Tsuchikawa S. (2010) Effect of Crosscutting Crooked Sugi (Cryptomeria japonica D. Don) Logs on Sawing Yield and Quality of Sawn Lumber. Forest Products Journal, 60(3):244-248

240


Increased sawing yield by thinner saw blades and adapted green target sizes Flodin, Jens1 and Grönlund, Anders2 1

SP Technical Research Institute of Sweden (SP Trätek) 2

Luleå University of Technology, Skellefteå Sweden

ABSTRACT An issue of growing interest for today’s sawmills is the utilization of the raw material as approximately 70 % of the sawmill costs can be derived from raw material costs. Hence it is very important for sawmills to obtain highest possible yield. Saw blade thickness and green target size are important parameters that affects the sawing yield. Modern saw lines often have high productivity with high feed speeds, thick saw blades but also often very good measurement accuracy of the sawn timber. High precision in measurement accuracy makes it possible to make a very precise adaptation of the green target sizes dependent of where in the cross section the piece is sawn. The purpose of this study was to highlight the positive effect on the sawing yield that a decreased saw blade thickness along with adapted green target sizes could bring. The sawing simulation software Saw2010 has been used along with 1596 true shape Norway spruce logs collected at the saw intake of a sawmill in northern Sweden. Three multiple ex sawing patterns, that are sawn at a sawmill in northern Sweden, have been used in the study. The results show a decrease in saw blade thickness by 1 mm can increase the sawing yield by more than 3 %-units in multiple-ex sawing patterns. If green target sizes are adapted to cross sectional position together with a decreased saw blade thickness the sawing yield could increase by more than 5 %-units dependent on the sawing pattern.

241


INTRODUCTION An issue of growing interest for today’s sawmills is the utilization of the raw material as approximately 70 % of the sawmill costs can be derived from raw material costs. Hence it is very important for sawmills to obtain highest possible yield. Saw blade thickness and green target size are important parameters that affects the sawing yield. Which saw blade thickness that is used depends on saw line feed speed, targeted sawing accuracy and risk for catastrophic failure. Modern saw lines often have high productivity, high feed speeds and thick saw blades with negative impact on sawing yield as a result from this. At the same time modern saw lines often have very good measurement accuracy of the sawn timber. The shrinkage of boards occurs whenever wood is dried below the fiber saturation point (roughly 30% moisture content). How much a specific board will shrink during drying depends of course to which moisture content it is dried but also on where in the log cross section the board is cut i.e. how much of the shrinkage is tangential and how much is radial shrinkage in the specific board [1]. The purpose of this study was to highlight the positive effect on the sawing yield that a decreased saw blade thickness along with adapted green target sizes could bring.

MATERIAL AND METHOD In this study the sawing simulation software Saw2010 has been used in order to investigate how much the sawing yield potentially could be raised by using thinner saw blades in both the first- and second saw along with adapted green target thickness on the centre yield dependent on its distance from the pith. Figure 1 shows a screen dump from the simulation software [ 2].

Figure 1 Screen dump from Saw 2010

242


Saw2010 gives possibility to set many different parameters that affect the sawing performance. These are for example curve sawing ability, saw blade thickness, green target size, sideboard optimization and rotation- and positioning error. It’s also possible to execute scripts in order to simulate a large number of logs with different settings. In this study three different scenarios have been simulated. 1. Scenario 1, with 4,4 mm saw kerf and no differentiation in green target thickness 2. Scenario 2 with 3,4 mm saw kerf and no differentiation in green target thickness 3. Scenario 3 with 3,4 mm saw kerf and adapted green target thickness in the 2nd saw dependent on distance from pith Three different multiple ex sawing patterns, that are used at a sawmill in northern Sweden, have been used as example when simulating each scenario. Table 1 shows the sawing patterns with nominal and green target sizes used in the simulations. Gray areas show the green target size settings for scenario 1 and 2 while white areas show the adapted green target sizes used in scenario 3. Table 1: Sawing patterns for second saw with nominal- and green target dimensions. Gray areas show the green target size settings for scenario 1 and 2 while white areas show the adapted green target sizes used in scenario 3. Scenario Nominal Thickness No DS_POST 1 , 2 4ex‐21x95 16/21/21/21/21/16 6ex‐23x145 1 , 2 22/23/23/23/23/23/23/22 6ex‐27x200 1 , 2 22/27/27/27/27/27/27/22

Green Target Thickness DS_POST_GS

ID_NAME

17.1 / 23.6 / 23.6 / 23.6 / 23.6 / 17.1 23.3 / 26.9 / 26.9 / 26.9 / 26.9 / 26.9 / 26.9 / 23.3 23.3 / 32.2 / 32.2 / 32.2 / 32.2 / 32.2 / 32.2 / 23.3

4ex‐21x95

3

16/21/21/21/21/16

17.1 / 23.2 / 23.6 / 23.6 / 23.2 / 17.1

6ex‐23x145

3

22/23/23/23/23/23/23/22

23.3 / 25.6 / 26.2 / 26.9 / 26.9 / 26.2 / 25.6 / 23.3

6ex‐27x200

3

22/27/27/27/27/27/27/22

23.3 / 30.6 / 31.3 / 32.2 / 32.2 / 31.3 / 30.6 / 23.3

For each scenario sawing simulation have been performed on 1596 true shape Norway spruce logs collected at the saw intake of a sawmill in northern Sweden. In this study there have been no rotation- and positioning error involved while other parameters, except saw blade thickness and green target size, have been kept constant between scenarios in order to more clearly see the influence on sawing yield from thinner saw blades and adapted green target thicknesses. The allowed sideboard widths used in the simulation was 75/100 mm for 16 mm sideboards and 75/100/125/150 mm for 22 mm sideboards.

243


RESULTS AND DISCUSSION Figure 2 shows the simulation results from the different scenarios as yield curves. The yield is the ratio between the nominal sawn timber volume and the entire log volume under bark.

Figure 2 Simulation results presented as yield curves From figure 2 it’s noticeable that there is a difference in sawing yield between scenario 1 and scenarios 2 and 3. Figure 3 shows the difference between scenario 1 and scenarios 2 and 3 where scenario 1 is the zero level.

244


Figure 3 Difference in sawing yield between different simulation scenarios From figure 3 one can observe that there is a maximum yield difference of approximately 3% for all sawing patterns by using 1 mm thinner saw kerf. For patterns 23x145 and 27x200 there is a maximum yield difference of approximately 5% when using 1 mm thinner saw kerf together with adapted green target sizes. This increase to 5% is however absent in pattern 21x95. A likely explanation for this is that there is less number of centre yield pieces in the pattern, i.e. less cuts to make, along with the fact that there is a smaller amount of alternative sideboard widths to use.

CONCLUSIONS The results from this study show that there is a good potential to raise sawing yield by using thinner saw blades along with adapting green target sizes to the sawn pieces position in the logs cross section.

REFERENCES 1.

Grönlund, A. ; Flodin, J. ; Wamming, T. 2009. Adaptive control of green target sizes. In Proceedings of 19th International Wood Machining Seminar, Nanjing, China Oct 21 – 23, 2009.

2.

Nordmark, U. 2005. Value Recovery and Production Control in the Forestry – Wood Chain using simulation Technique. LTU, Doctoral Thesis 2005:21; ISSN:1402-1544.

245


246

The Sawing Process Poster Presentations


Sawing Patterns for Sugi Large Logs Using at Sawmills in Japan Murata Kohji, Ikami Yuji, Matsumura Yukari Forestry and Forest Products Research Institute Matsunosato 1, Tsukuba, Ibaraki, JAPAN

ABSTRACT According to the mature of sugi (Cryptomeria japonica D. Don) planted forests, the main supply of sugi planted logs is expected to shift from small and middle logs to large logs in the near future. Common quality logs will be dominant and high quality logs will be a little in the near future supply of sugi large logs, because an adequate forest operation has not been carried out in most sugi planted forests. Therefore, the effective utilization of common quality sugi planted large logs will be strongly requested in Japan in the near future. Since 77 percent of sugi log is converted into sawn lumber now and most of the common quality sugi planted large logs will be also converted into sawn lumber in the near future, it is important to grasp the state of sawing sugi large logs for the effective utilization. In this study, we investigated the sawing patterns for sugi large logs at sawmills in several areas in Japan, in order to grasp the state of sawing sugi large logs. We also investigated the price of sugi large logs, the size and price of the main sawn lumber, and the sale area.

INTRODUCTION In Japan, coniferous trees had been planted studiously after the Second World War, in order to repair ruined forests and in order to increase the amount of lumber production for wooden housing. Consequently, the planted forest area had reached 10 billion hectares in 1970s and the planted forest area of 10 billion hectares has been kept until now. The major planted species are sugi (Cryptomeria japonica D. Don) and hinoki (Chamaecyparis obtuse Endlicher). Their shares in the planted forests area are 44% and 25% and those in the planted forest stock are 57% and 22%, respectively. Sugi has been the most representative planted species in Japan. Since the sugi logs from planted forests have been used for columns in the Japanese traditional house, the cutting age of sugi planted forest was set as 40-50 years depending on the planted sites. Many sugi planted forest has not been cut at the cutting age, because forest owners would not like to cut their forests. The sugi planted forests has matured recently. According to the mature of sugi planted forests, the main supply of sugi planted logs is expected to

247


shift from small (8-13 cm in diameter) and middle (14-28 cm in diameter) logs to large (more than 30 cm in diameter) logs in the near future. Common quality logs will be dominant and high quality logs will be a little in the near future supply of sugi large logs, because an adequate forest operation has not been carried out in most sugi planted forests. Therefore, the effective utilization of common quality sugi planted large logs will be strongly requested in Japan in the near future. Since 77 percent of sugi log is converted into sawn lumber now and most of the common quality sugi planted large logs will be also converted into sawn lumber in the near future, it is important to grasp the state of sawing sugi large logs for the effective utilization. In this study, we investigated the sawing patterns for sugi large logs at sawmills in several areas in Japan, in order to grasp the state of sawing sugi large logs. We also investigated the price of sugi large logs, the size and price of the main sawn lumber, and the sale area.

METHOD We investigated the sawing patterns for sugi large logs at each 2 sawmills in Akita Prefecture, Wakayama Prefecture, Tottori Prefecture, and Miyazaki Prefecture. We also investigated at 1 sawmill in Gifu Prefecture. Therefore the total number of investigated sawmills were nine. Their locations are shown in Figure 1. We visited these seven sawmills and investigated the sawing patterns for sugi large logs, the price of sugi large logs bought by them, the size and price of the sawn lumber produced by them, and their sale area (actual in 2009).

A,B

◎

I E,F ◎ ◎

◎ ◎ ◎ ◎

G

C

D

H

Figure 1 Location of the investigated sawmills.

RESULTS AND DISCUSSION Logs consumed at the investigated sawmills Table 1 shows the amounts of log consumption at the investigated nine sawmills. The amounts of log consumption were from 1,800 m3/year to 20,000 m3/year. The shares of sugi logs were from 70 % to 100 %. Miyazaki Prefecture, locates the southern part of Kyusyu Island, is logging the largest amount of sugi logs in Japan and Akita Prefecture, locates the northern part of Honsyu Island, is logging the second largest amount of sugi log. The shares of sugi logs at sawmills in these two prefectures were greater than those in the other prefecture. Hinoki does not distribute in Akita

248


Prefecture, so that himoki was not sawn at two sawmills there. Table 1 Amounts of log consumption at the investigated sawmills. Sawmill A B C D E F G H I

Location Noshiro, Akita Noshiro, Akita Kamitomita, Wakayama Kushimoto, Wakayama Chhizu, Tottori Chhizu, Tottori Hinokage, Miyazaki Hyuga, Miyazaki Takayama, Gifu

Log consumption (m3/year) 4,000 20,000 10,000 3,000～3,500 3,120 1,800 34,000 15,000 17,000

Species sugi(98%), hiba(2%) sugi(100%) sugi(85%), hinoki(15%) sugi(70%), hinoki(30%) sugi(75%), hinoki(15%), matsu(10%) sugi(80%), hinoki(20%) sugi(95%) , hinoki(5%) sugi(100%) sugi(100%)

Table 2 shows the diameters and lengths of logs consumed at the investigated sawmills. The shares of sugi large (more than 30 cm in diameter) logs were more than 30% at six investigated sawmills. The lengths of sugi logs were mainly 3 m and 4 m, despite two sawmills in Akita Prefecture. Traditionally 3.65 m long logs have been produced in the northern part of Honsyu Island and 3.65 m long logs were produced the most in Akita Prefecture. Therefore, the shares of 3.65 m long logs occupied 80 % at two sawmills in Akita Prefecture. Table 2 Diameters and lengths of logs consumed at the investigated sawmills. Sawmill

Species

A

sugi

-18cm(20%), 20-30cm(50%), 30cm-(30%)

3.65m(80%), 4m(20%)

sugi

14-22cm(34%), 24-30cm(34%), 30cm-(32%)

3.65m(80-85%), 4m(15-20%)

-16cm(20%), 24-30cm(30%), 30cm-(30%)

4m（major）, 3m, 5m

B C D E F G

sugi hinoli sugi hinoki sugi hinoki sugi hinoki sugi hiniki

Log diameter

Log length

18-30cm(60%), 30-40cm(30%), 40-60cm(10%) -30cm(90%), 30cm-(10%) 14-18cm(10%), 20-30cm(70%), 30cm-(20%) -18cm(70-80%), 30cm-(20-30%) -30cm(10%), 32-36cm(50%), 38-48cm(40%) 13-18cm(50%), 20cm-(50%) 14-24cm(70%), 26-30cm(15%), 30cm-(15%) -30cm

3m(10%), 4m(60%), 4m-(30%) 4m 3m, 4m 3m, 4m

3m, 4m, (special order: 6-12m, 16m)

H

sugi

24-28cm(55-60%), 30cm-(40-45%)

3m(15%), 4m(80%), 6m, 7m, 8m

I

sugi

14-18cm(24%), 20-24cm(43%), 26-30cm(28%), 32cm-(5%)

3m(50%), 4m(50%)

Table 3 shows the log supplier for the investigated sawmills. The investigated sawmills got logs from loggers, log markets, forest cooperatives, forest owners, and own forests. The number of log markets is more in the western and southern part of Japan than in the eastern and northern part of Japan, so that the rate of the logs got from log markets are higher at sawmills in the western and southern part of Japan than at those in the eastern and northern part of Japan. Two sawmills purchased stumpage. Four sawmills got logs within 50 km away and two sawmills within 60 km. The price of large logs purchased by the investigated sawmills is shown in Table 4. The log price depends on the species, diameter, length, and quality of logs. In general, the log price trends to increase with an increase in log quality, if the quality is same. However, the log price of sugi decreased with an increase in log diameter at Sawmill H. The sugi common quality large logs was less demand than sugi common quality middle logs in the area where Sawmill H locates, so that the price of sugi large log was less than that of sugi middle log. Common or low quality logs are sold and bought by a lot and high quality logs are sold and bought by each log in Japan.

249


Table 3 Log supplier for the investigated sawmills. Sawmill

Supplier

Dicetance

A

log market (20%), forestry cooperative (80%)

within 50km

B

logger (30%), forestry cooperative (35%) log market (35%)

within 60km

C

log market (50%), logger (50%)

within 40km

D

log market (5%), forest ower (65～75%) own forest (20～30%)

within 60km

E

log market (90%), forest cooperative (10%)

within 50km

F

log market (100%)

within 50km

G


within 50km

H


within 30km

I

log market (100%)

Stumpage buying

20%

95%

within 80-90 km

Table 4 Prices of large logs purchased by the investigated sawmills. Sawmill

Species

A

sugi hiba

B

sugi

C D E F G H I

sugi hinoki sugi hinoki sugi hinoki sugi hinoki sugi hinoki sugi

30

32

Log diameter (cm) 34 36 38 11,000-15,000 (high quality: 40,00-50,000) 80,000-100,000

11,500-12,000

unit:JPY/m3 40-

11,500-36,000 16,000-17,000

18,000-20,000 (60-70cm: 40,000-80,000) 50-60cm: 150,000-400,000 28,000-30,000 (high quality: more than 50,000) 35,000-45,000 20,000 50,000-60,000 200,000 11,000（straight: 12,000-13,000, warp: 8,000) 11,500

12,800～13,000

10,500

12,500-14,000 13,800-14,500

sugi

Sawing patterns at the investigated sawmills Figure 2 shows the examples of sawing pattern for sugi large log at Sawmill A. Basically, shokaku or hirakaku was sawn prior to the other lumbers. Shokaku is a lumber whose dimension is more than 7.5 cm in thickness and width and its cross section is square. Shokaku is mostly used for columns in the Japanese traditional house. Hirakaku is a lumber whose dimension is more than 7.5 cm in thickness and width and its cross section is rectangular. Hirakaku is mostly used for beams the Japanese traditional house. The lumber for fixtures was sawn from high quality side boards. The rough sawn size of shokaku was 135 by 135 mm (finished size: 120 by 120 mm) or 120 mm by 120 mm (finished size: 105 by 105 mm) and that of hirakaku was 120 mm by 170, 200, 230, 260, 290, and 320 mm (finished size: 105 by 150, 180, 210, 240, 270, and 300 mm). The thicknesses of rough sawn lumbers for fixtures were 33, 39, and 48 mm (finished thickness: 30, 36, and 45 mm).

250


135×135mm 120×120mm

135×135mm 120×120mm

Figure 2 Examples of sawing pattern for sugi large log at Sawmill A.

Figure 3 shows the examples of sawing pattern for sugi large log at Sawmill B. Two sawing patterns shown in Figure 3 were used for high quality logs. In the case of 36 cm diameter log, 45 mm thick lumber for fixtures was sawn at the highest quality log face at first, then plural pieces of 134 mm by 134 mm rough sawn shokaku (finished size: 120 by 120 mm) were sawn. In the case of 46 mm diameter log, 39 mm thick quarter sawn lumber (finished thickness: 36 mm) was sawn prior to the other lumbers and plural pieces of shokaku, scantlings, and boards were sawn.

134×134mm

39mm thick quarter sawn lumber

fixtures 134×134mm

134×134mm 45×125mm

rafter

134×134mm

boards

134×134mm rafter 134×134mm

46cm diameter log (high quality)

36cm diameter log (high quality)

Figure 3 Examples of sawing pattern for sugi large log at Sawmill B.

48×190mm

Figure 4 shows the example of sawing pattern for sugi large log at Sawmill C. A piece of 115 or 130 mm thick rough sawn hirawari (finished thickness: 105 or 120 mm) was 115mm thick sawn from the center part and 39-48 mm thick unedged 130mm thick hirakaku planks (finished thickness: 36-45 mm) were sawn from the side parts. After drying, the unedged planks were resawn into scantling (for stud, brace, and fixtures) and their Figure 4 Example of sawing pattern finished sizes were 36 and 45 mm by 45, 55, 60, 90, and for sugi large log at Sawmill C. 105 mm. Figure 5 the examples of sawing pattern for sugi large log at Sawmill D. In the case of 20-30 cm diameter log, 30-40 mm thick unedged planks were sawn from the side parts and a piece of 135 mm thick rough sawn hirakaku (finished size: 120 mm) was sawn from the center part. The finished sizes of the unedged planks were 24, 30, 34, 36, and 40 by 12-270 mm (random width). In the case of 30-50 cm diameter log, 10-36 mm thick boards were sawn from the side parts at first, then 45-60 mm thick boards were sawn, and finally a piece of 135 by 165-345 mm hirakaku and 45 mm or 30-36 mm thick lumber were sawn from the center part. The 45-60 mm thick boards were resawn into scantlings (for fixtures, 45 by 12 and 15 mm and/or members of stair, 45-60 by 300-400

251


mm). The finished sizes of hirakaku were 120 by 150, 180, 210, 240, 270, 300, and 330 mm. Before they had been sawing warikaku (shokaku without pith) at Sawmill D but they quitted sawing warikaku because of the laborious procedure.

45-60mm thick

10-36mm thick

135×165-345mm

135×135mm or 135×165-255mm

45-60mm thick

10-36mm thick

45mm or 30-36mm thick

30,34,36,40mm×120-270mm

30-50cm diameter log

20-30cm diameter log

Figure 5 Examples of sawing pattern for sugi log at Sawmill D.

Figure 6 shows the examples of sawing pattern for sugi large log at Sawmill E. These two sawing patterns were applied to 40 cm diameter logs according to the log quality. In the case of the top in Figure 6, 2 pieces of rough sawn hirakaku (135 by 225, 255, and 285 mm) and 18 mm thick wide boards were sawn. In the case of the bottom in Figure 6, rough sawn scantlings (48 by 135 mm), clear wide boards, and 36 mm thick rough sawn planks were from the not free and high quality log face, and 18 mm thick rough sawn wide boards and a piece of rough sawn hirakaku were sawn from the lower quality log face. The finished size of the rough sawn scantlings, 18 mm thick wide boards, 36 mm thick planks, and hirakaku were 45 by 120 mm (for fixtures), 15 by 150, 180, and 210 mm (for rakes), 30 by 180 mm (for flooring), and 120 by 210, 240, and 270 mm, respectively.

48×135mm

225 135×255mm 285

36mm thick

225 135×255mm 285

knot free, high quality

36mm thick

225 135×255mm 285

inadequate for fixtures

36mm thick

18mm thick

18mm thick

Figure 6 Examples of sawing pattern for sugi large log at Sawmill E.

Figure 7 shows the example of sawing pattern for sugi large log at Sawmill G. Thirty-five mm thick planks or 50 mm thick flitches were sawn from the side parts and 2 pieces of rough sawn hirakaku (118 by 170 and 200 mm) were sawn from the center part. The finished sizes of hirakaku were 105 by 150 and 180 mm. The 35 mm thick planks and 50 mm thick flitches were resawn into 30 by 105 mm (for stud) and 45 by 45 mm and 45 by 60 mm (for rafter), respectively. Figure 8 shows the example of sawing pattern for sugi large log at Sawmill H. At first, 15 mm thick boards are sawn from the side parts, then 33-63 mm thick rough sawn planks are sawn, finally 115 or 130 mm thick rough sawn hirakaku is sawn from the center part. Thirty-three to sixty-three mm thick rough sawn planks were resawn into 30, 40, 45, and 60 mm thick scantlings. The finished sizes of hirakaku were 105 and 120 by 150 and 180 mm.

252


33mm 43mm 48mm 63mm thick

Figure 8 Example of sawing pattern for sugi large log at Sawmill H.

Figure 9 shows the example of sawing pattern for sugi large log at Sawmill I. The break down was carried out with a twin bandmill with an auto feed carriage at this sawmill, so that the sawing pattern was relatively simple. Thirty-five mm thick rough sawn planks were sawn from the side parts and a piece of 135 mm thick rough sawn hirakaku was sawn from the center part. The 35 mm thick rough sawn planks were finished into 27 by 120 mm boards. The finished sizes of the 135 mm thick rough sawn hirakaku were 120 by 180, 210, 240, and 270 mm.

135×225mm

35mm thick

Figure 7 Example of sawing pattern for sugi large log at Sawmill G.

115,130×220,250mm

118×170,200mm

33mm 43mm 48mm 63mm thick

35mm thick

118×170,200mm

35, 50mm thick

35, 50mm thick

15mm thick boards

Figure 9 Example of sawing pattern for sugi large log at Sawmill I.

Sawn lumber at the investigated sawmills Table 5 shows the sizes of the main lumber from large logs at the investigated sawmills. Hirakaku was the main lumber from sugi large logs at 8 sawmills of 9 investigated sawmills. Hirakaku is a larger size than the other lumber, so that it is easy to be sawn from large logs. The lumbers for fixtures and furniture members were prior to the other lumbers, because they could be sold at a high price. Table 5 The sizes of the main lumber from large logs at the investigated sawmills. Sawmill

Species

A

H

sugi hiba sugi sugi hinoki sugi hinoki sugi hinoki sugi hinoki sugi hinoki sugi

I

sugi

B C D E F G

30

Log diameter (cm) 34 36

32

38

10.5×10.5, 12.0×12.0, 10.5,12.0×15.0- (3cm increment) 10.5×10.5, 12.0×12.0, 4.5, 5.5×4.5 ,5.5

3.0 thick 10.5, 12.0×15.0-

4.5, 6.0×30.0, 12.0×15.012.0×12.0, 4.5,6.0×30.0、2.4,3.0,3.4,3.6,4.0×12.0-30.0 12.0×18.0, 21.0, 24.0, 27.0, 1.5, 3.0×18.0 3.0×18.0 12.0×21.0-, 1.5×12.0 10.5, 12.0×15.0-, 3.0×10.5, 4.5×4.5, 6.0 10.5 ,12.0×21.0, 24.0, 2.8×19.0, 3.4×17.0, 1.5×4.5-15.0

40～

unit: cm Remarks priority: fixtures priority: fixtures priority: fixtures priority: fixtures priority: fixtures priority: furnuture member priority: fixtures

12.0×18.0-2.7×12.0

Table 6 shows the selling price of the main lumber at the investigated sawmills. The lumber price depends on the species, size, and quality of lumber. In Japan, the lumber which is without knot on its surface and has good fine grain is sold at a high price. For example, the price of the common grade warikaku was 55,000-60,000 JPY/m3, however, that of the high grade was 80,000-140,000 JPY/m3 at Sawmill A. Also the lumber price is different depending on the distinct even if the

253


lumber grade is same. Table 6 The selling price of the main lumber at the investigated sawmills. Sawmill Species Selling price of the main lumber (JPY/m3) sugi scantling (G, C): 25,000-30,000, scantling (KD, C): 45,000-50,000、warikaku (C): 55,000-60,000 A hiba lumber for fixtures: 120,000-150,000, warikaku (H): 80,000-140,000 shokaku (G, B): 31,000, shokaku (KD, B): 48,000, scantling (G, C): 36,000, scantling (KD,C): 43,000 B sugi warikaku (AD, C): 46,000, warikaku (AD, H): 100,000-, quarter sawn boards for fixtures (AD): 100,000 sugi hirakaku (KD, C): 70,000-85,000 C hinoki sugi shokaku (C): 40,000-45,000, hirakaku (C): 45,000-80,000, table counter: 50,000-800,000 D hinoki shokaku (C): 50,000-70,000, hirakaku (C): 70,000-120,000, table counter: 100,000-2,000,000 sugi hirakaku: 55,000-65,000, boards for interior use: 150,000-200,000 E hinoki shokaku: 75,000-90,000 sugi shokaku (KD, B), 60,000-70,000, hirakaku : 75,000, warikaku rikaku (including 4 sides clear): 200,000 F hinoki sugi hirakaku (KD, C): 52,000, shokaku (KD, C): 47,000, stud (KD, C): 45,000 G hinoki board (G, C): 40,000, hirakaku (G, C): 30,000-36,000, scantlings (G, C): 33,000-40,000、 H sugi scantlings (G, H): 60,000-70,000 I

sugi

shoksku and hirakaku (KD, C): 52,000、stud (KD, C): 41,000, thin boards (G, C): 18,000

G: green, AD: air dried. KD: kiln dried. C: common grade, H: high grade for interior lumber, B: boxed pith

Table 7 shows the rate of shipping inside and outside Table 7 The rate of shipping inside the prefecture. The rates of shipping outside the and outside the prefecture. prefecture ranged from 30 % to 100 %. The rates of Sawmill Inside Outside 70% (Niigata, Toyama, Kanto) A 30% shipping outside the prefecture was higher at the 70% (Saitama, Tokyo, Kanagawa) B 30% 30% (Kyushu, Kinki, Tokyo) C 70% sawmills in Akita Prefecture and Miyazaki Prefecture 65% (Kanto, Mie, Kyushu) D 35% 60% (Okayama, Hyogo) E 40% than in the other prefectures. These two prefectures 30% (Okayama, Tokyo, Aichi) F 70% 100% (Fukuoka, Nagasaki, Aichi) G are ones of the prefectures in which the amount of 50% (Aichi, Osaka, Oita, Fukuoka) H 50% lumber manufactured is larger. They are 80% (OEM) I 20% manufacturing sawn lumber not for inside the prefecture but for the high consumption areas such as Kanto, Tokai, and Kinki. Although the OEM of sawn lumber is very rare in Japan, 80 % of the lumber sawn at Sawmill I was OEM. They were sold as the products of a famous sawmill.

CONCLUSION We investigated the sawing patterns for sugi large logs at sawmills in several areas in Japan, in order to grasp the state of sawing sugi large logs. The sawing patterns were different depending on log diameter and quality and the area. Hirakaku and the lumbers for fixtures were mainly sawn at the investigated sawmills. The lumbers for fixtures can be sawn from the higher quality logs but not from the common quality logs. It is estimated that the common quality logs will be dominant and the high quality logs will be a little in the near future supply of sugi large logs, because an adequate forest operation has not been carried out in most sugi planted forests. Since we will not be able to expect the supply of large amounts of sugi higher quality large log, hirakaku might be one of the options.

This study was funded by The Japan Forestry Association in 2010.

254


Wood Processing Capacity of Sawmills and Carpentry Workshops in Ghana Appiah-Kubi, Emmanuel1, Adom-Asamoah, Mark2, Frimpong-Mensah, Kwasi3 and Tekpetey, Stephen Lartey1 1

2

Forestry Research Institute of Ghana (CSIR-FORIG), Kumasi, Ghana Civil Engineering Department, Kwame Nkrumah University of Science and Technology, Ghana 3 Department of Wood Science & Technology, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana

ABSTRACT In Ghana, the exploitation of timber is limited to a few of the over 300 known species. Majority of the species are not being utilized because their properties (including physical, mechanical and machining properties) are not known. Due to this, sawmills hardly process these lesser used species. To avoid the overexploitation of commercially known species the use of lesser known ones is inevitable. The objective of the study was to assess the capacity of the sawmills and carpentry workshops in terms of their machinery to process lesser used timber species for efficient utilization. Eight (8) of the sawmills in Kumasi, Ghana were randomly selected and questionnaires were administered and interviews conducted. Forty-five (45) carpenters granted interview and responded to questionnaires that were administered. The sawmills have the needed cutting and processing machinery for producing lumber from commercially known species but not the lesser known ones. Some sawmills expressed difficulty in sawing some lesser used timber species due to their extreme hardness and smaller diameter sizes even though they possess adequate strength for utilization. Most of the carpentry workshops (60%) use only simple hand tools for processing which makes the utilization of these lesser known species in construction very difficult. Advanced technology and techniques in processing which are cost effective needs to be developed for the processing of these species for efficient utilization in order to reduce the overexploitation of the commercial (traditional) timber species.

INTRODUCTION Majority of the tree species in Ghana are not being utilized because most of their properties are not known. The exploitation of timber is limited to a few of the over 300 known species. Until now less than about 100 tree species are seriously commercially utilized and the rest unexploited [1]. There are a number of timber companies, which produce timber to the required sizes in commercial quantities [2] but most of them hardly process the lesser used species. To avoid the overexploitation of commercially known species the use of lesser known ones is inevitable. The use of lesser used species for construction require an assessment of the state of the sawmill industries and their

255


readiness to process these species. The Timber Industry Development Division (TIDD) of the Forestry Commission of Ghana has norms and regulations for sawmill operation [3]. An important regulation is to have the necessary machinery before one can apply for a permit to operate a sawmill. Moreover, sawmills can either apply for a working area to obtain the logs from or buy from loggers or do both. Smaller mills produce timber for the local market but sawmills which produce largely for export are required to produce 20% for the local market and are expected to submit documents to prove this. The objective of the study was therefore to assess the capacity of the sawmills and carpentry workshops in terms of their machinery to process lesser known timber species for efficient utilization.

MATERIALS AND METHODS Eight (8) sawmills were sampled in Kumasi for the study. Most of them produce lumber and other products mainly for export. Structured questionnaires were administered and semi-structured interviews were also conducted. A descriptive statistical method was used to analyse the responses and the results presented in tables and graphs. Several carpentry shops and carpenters were visited in Kumasi to ascertain their capacity in working with the lesser used species. Forty-five (45) carpenters granted interview and responded to questionnaire that were sent. The study was conducted in the Kumasi (6[degrees] 54’N 1[degrees] 35’E) metropolis which has a proportionally large number of timber processing firms in Ghana [2].

RESULTS AND DISCUSSION Capacity of Sawmills The outcome of the survey is based on the responses of the eight sawmills who granted interview and responded to the questionnaire. Table 1: Type of machinery used by Sawmills Machinery No. of Sawmills Percentage Cutting/processing machines 8 100 Kiln Dryer 2 25 Moving machines 3 37.5

(%)

Table 1 shows that all the 8 sawmills have the various cutting and processing machinery such as band mills, cross cut saws, edgers, planners, rippers etc. In Ghana, lesser used species (tropical hardwoods are usually processed with either stellite-tipped or swage-set cutters [4]. Three of the sawmills which represent 37.5% have moving machines such as cranes, lifts, and forklifts for moving and transporting heavy materials. Only two (25%) of the sawmills have kiln dryers. The others either air dry or kiln-dry their wood at the sawmills with Kiln dryers. The lack of kiln dryers in most of the sawmills affect their production capacity since they have to wait for longer periods to air-dry or pay huge sums for kiln-drying their wood in other sawmills before delivery to clients.

256


Years of Usage 0% of Machinery 0% 25% < 1 year 1-5 yrs 5-10 yrs >10 yrs 75%

Figure 1: Chart showing years of usage of Sawmills’ machinery Table 2: Rate of Breakdown of Machinery Breakdown of Machines Frequency Percent (%) Very often (daily/weekly) 1 12.5 often (monthly) 1 12.5 Not often (3-6 months) 1 12.5 Hardly (yearly or more) 5 62.5 Total 8 100 Seventy-five percent (75%) of the sawmills have used their machinery for over 10 years and 25% of them had used their machinery between 5 and 10 years (Figure 1). None of the sawmills visited had a processing machine purchased within the last five years. The sawmills also indicated that most of their machines were purchased brand new except for some brands which were slightly used. Most of the machinery used by the sawmills were imported from Germany, England and Italy [5]. Five of the sawmills (62.5%) indicated that their machines hardly breakdown (yearly or over) and only one (1) sawmill indicated that their machines breakdown very often (daily or weekly) causing delays in meeting contract deadlines (Table 2). This explains why most of the machines had been in use for over 10 years without replacement (Figure 1) and could be as a result of the installation of new machinery at the sawmills between 1980 and 1985 as part of a major rehabilitation work to increase the sawmills’ efficiency [5]. Table 3: Processing of lumber for local market Reasons Frequency Percentage (%) Easy transport 0 0 Inexpensive handling 2 25 TIDD regulation 5 62.5 Ready market 1 12.5 Total 8 100 All the eight sawmills (100%) responded that they produce lumber for the local market. They gave reasons as indicated above (Table 3). Five (62.5%) of them said they produce for the local market because it is regulation from the Timber Industry Development Division of the Forestry Commission of Ghana. Twenty-five percent (25%) said they process for the local market because it easy and inexpensive in handling products to the local market compared to the export market. One sawmill

257


indicated that the local market is a readily available market even if there are no foreign contracts. The results above indicate that the sawmills are sending lumber to the local market because it is a regulation by the Government (TIDD). Table 4: Processing of lumber for export Reasons Frequency Percentage out of 8 (%) Foreign exchange 5 62.5 12.5 Profit 1 Expensive for local 3 market 37.5 Local market unavailable 3 37.5 All the sawmills (100%) again indicated that they produce lumber for export. Five of the sawmills (62.5%) produce for the export market because of the foreign exchange and 37.5% of the sawmills said their products are expensive for the local market. This is because their production cost is high and they spend a lot of money maintaining their machines for production. The local market is also not ready to pay for the high cost of production so the sawmills export for foreign exchange in order to make profit. All the sawmills (100%) indicated that they produce 50mm x 50mm and 50mm x 100mm sizes of lumber for the local market. Only one company (12.5%) produce 100mm x 150mm lumber size for the local market and none of the firms currently produce 150mm x 150mm and 150mm x 200mm lumber sizes for the local market. The sawmills attributed this to the demand of the local market. There is no demand for structural size lumber such as 100mm x 150mm, 150mm x 150mm and 150mm x 200mm so the Sawmills do not produce them. Capacity of Carpentry Workshops Sixty percent (60%) of the carpenters use simple hand tools such as hammers, chisels, spirit levels, planes etc. in their workshops while 40% of the respondents indicated that they use both simple hand tools and machines such as planners, cross cut saws, table saws, circular saws etc. This means that carpenters are more familiar with the use of simple hand tools than the use of machines. Table 5: Type of tools and equipments used by carpenters Type of Equipment Frequency Percent (%) Simple hand tools 27 60 Machines 0 0 Hand tools & Machines 18 40 Total 45 100 Table 6: Rate of repairs or replacement of tools and machines Repairs and Replacement Frequency Percent (%) Very often (daily/weekly) 0 0 often (monthly) 13 29 Not often (every 3-6 months) 23 51 Hardly (yearly or more) 9 20 Total 45 100

258


Twenty-three (51%) of the respondents indicated that their machines or tools do not breakdown often. They usually repair or replace them every 3 – 6 months. Thirteen (13) carpenters (29%) indicated that their machines breakdown often. They repair or replace some hand tools monthly. However, 20% said they hardly repair machines or replace tools. They do such yearly.

Availability of Spare Parts for Repairs 50

42

40 30 20 10

3

0

0 Available (locally)

Available (imported)

Unavailable

Figure 2: Chart showing the availability of spare parts for the repair of machinery Forty-two carpenters (93%) indicated that spare parts for the repairs of machinery are available in the local market. All the carpenters also said that they do not have maintenance departments in their workshops but usually employ the services of engineers at Suame Magazine, Kumasi when breakdowns occur. Table 7: Grades of lumber in the local market Grades FAS No.1 C&S No.2 C&S Total

Frequency 0 9 14 23

Percent (%) 0 39 61 100

Fifty-one percent (51%) of the carpenters know about the grading and different grades of wood. 22 of the respondents, representing 49% do not know anything about the various grades of wood. They however said they have their own way of selecting and grading wood for their works. Out of the 23 carpenters who know about the various grades, 61% uses the No. 2 C&S (which is the least grade) and 39% uses the No.1 C&S. According to the respondents, the FAS (i.e. First And Second) grade is not available in the local market because the sawmills export them always. Thus, the FAS can only be obtained from the export market and is very expensive.

259


Table 9: Reasons for using sawmill lumber Reasons If Yes Frequency Percentage Guarantee of wood 4 44.4 quality & grade Wood is easy to work with 0.0 Client's preference 2 22.3 Others sources are Illegal 3 33.3 Proximity to Sawmill 0.0 100 Total 9 Table 10: Reasons for not using sawmill lumber Reasons If No Frequency Percentage Too expensive 29 80.6 Sawmills are far 2 5.6 Bureaucracy & Security checks 4 11 Difficulty in transport 1 2.8 Total 36 100 Eighty percent (80%) of the carpenters do not obtain their wood from sawmill whilst nine (20%) of them obtain their wood from sawmill. Out of the nine (9) respondents who obtain their lumber from sawmill, 44% choose to pay for the sawmill wood because of the guarantee of the wood grade and quality. This gives them good finish and so they are able to sell their products at higher prices. About 33% said it is illegal to obtain lumber from other sources such as chain saw operators and 22% obtain their wood from sawmill because some of their clients prefer sawmill wood (Table 9). The clients insist that they use only wood from sawmill and such clients are prepared to pay for higher cost. Twenty-nine (80.6%) out of the 36 respondents who do not obtain their wood from the Sawmills indicated that wood from sawmills are too expensive (Table 10). Eleven percent (11%) of the respondents do not buy wood from sawmills because of the bureaucracies at the sawmill and the several security checks. Before you obtain wood from a sawmill, you will have to place an order and make an advanced payment and provide proof of registration of your firm. The procedure is cumbersome and discourages carpenters from buying wood from sawmills. Ninety-one percent (91%) of the carpenters indicated their readiness to work on other lesser used timber species. Only 4 out of the 45 respondents said that they were not ready to work with other secondary species because they make their products with specific species of wood. Fifty-three percent (53%) of the carpenters have had no education while 47% of the carpenters indicated that they have had education of some sort. Eighteen (85.7%) of the educated carpenters had education up to the Junior High School Level. Two (2) carpenters had education up to Senior High School Level. None of the respondents have had education up to the tertiary level.

260


CONCLUSIONS AND RECOMMENDATIONS The sawmills in the Kumasi Metropolis have the needed machinery to process lesser used species for both export and local markets. They indicated their readiness to process for the local construction industry if the local customers are ready to pay for the cost of lumber from the sawmills. In Table 3, 62.5% of the sawmills said that they produce lumber for the local market only because it is regulation from the TIDD. Most of the sawmills (70%) were processing species such as Dahoma, Emire, Esia etc. which were lesser used and now being used as replacement for the premium species such as Milicia excelsa (Odum) and Khaya ivorensis (Mahogany) which are being overexploited. They however expressed difficulty in sawing or processing the lesser used species. Specific challenges include the blunting of saw blades and cutters. This increases the frequency with which saw blades and cutters are replaced. Advanced technology and techniques in processing which are cost effective therefore needs to be developed for the processing of these species for efficient utilization in order to reduce the overexploitation of the commercial (traditional) timber species.

ACKNOWLEDGEMENT Many thanks to the Swiss Development Agency (SDC) and the Swiss National Science Foundation, for providing financial support for this study. Appreciation also goes to the directors, managers and staff of the sawmills and various carpentry workshops who made their facilities available and granted interviews for the study.

REFERENCES 1. Ayarkwa, J. (2009) Timber technology handbook for researchers, polytechnic and university students. Classic Graphics Print, Kumasi, Ghana 2.

Baiden, B.K., Badu, E., Menz, F.S. (2005) Exploring the barriers to the use and potential of timber for housing construction in Ghana. Journal of Construction and Building Materials. June 2005. Gale Group, Farrington Hills, Michigan

3.

TIDD. (1998) Norms and regulations for operations of timber processing firms. Timber Industry Development Division, Forestry Commission, Ghana.

4.

Okai, R., Mitchual, S. J., Frimpong-Mensah, K. (2001) Sawing accuracy of Stellite-tipped and Swage-set bandsaw blades when sawing a tropical hardwood at a Ghanaian sawmill. Proceedings of the 15th International Wood Machining Seminar. California, USA. pp. 229 – 238.

5.

Okai, R. (1999) Technology auditing of wood processing industries in sub-saharan Africa: Case study of the sawmilling industry in Ghana. Paris, France. pp. 709 – 717.

261


Development of Sustainable Bamboo Industries in Ghana: The Factors That Will Interplay. Tekpetey, Stephen Lartey and Appiah- Kubi, Emmanuel Council for Scientific and Industrial Research-Forestry Research Institute of Ghana (CSIR-FORIG), Kumasi, Ghana

ABSTRACT

Bamboo species in Ghana are environmentally friendly resources that could be processed for use in building and bridge construction, flooring and furniture making at relatively low prices instead of dwindling tropical hardwood species. Their sustainable utilization and the development of viable bamboo industries in Ghana are vital to its socio-economic advancement especially in the wake of climatic concern and the depletion of commercial tree species in the tropical forest. A number of factors are essential for the sustainable development of bamboo industries in Ghana using our native sympodial bamboo species. Appropriate policies that attract investors, acceptance of bamboo products as appropriate alternative to timber products, increased awareness of the benefits of using bamboo among local populace, capital investment and knowledge of the technical suitability of local bamboo species for specific products are some of the factors that will interplay towards the development of bamboo industries in Ghana. Forest managers should therefore ensure sustainable extraction of bamboo resources whilst wood technologists devise appropriate techniques for diversifying bamboo products and processing bamboo species in countries where the resource abound.

INTRODUCTION

Bamboo has been used in South Eastern Asia especially China and India and many other countries for centuries. Over the past two decades most of the industrial developments on bamboo have occurred in Asia. China and India, for instance, have developed a lot of bamboo technologies including the mass propagation of bamboo, bamboo gasifiers and the manufacture of bamboo panel products. They also use bamboo extensively in pulp and paper making and the trade in bamboo products keeps increasing. Industrial products such as bamboo flooring, sheet products (plywood, MDF), textile, charcoal and vinegar are being exported by china. The Philippines, although a small country with only about 50,000 hectares of bamboo plantations, is also reported to be the fourth exporter of bamboo and rattan products. Bamboo furniture exports from the Philippines have found a niche in the high-end market and are steadily growing [1]. Bamboo species grows in many areas of the world including Asia, Oceania & the Pacific, America and Africa and the increasing population and incomes around the world will further increase demand for forest resources and products especially bamboo products.

262


In Ghana, bamboo is underutilized and its use is limited to raw culms and splits for fencing and low- income housing in many deprived communities [2]. This is due largely to the lack of knowledge about the technical properties of native bamboo species and poor processing techniques [3]. Nevertheless, the creation of a sound bamboo industry would help ease the pressure on Ghana’s natural forests and provide alternative livelihood for people near the resource.

This paper will therefore review and highlight the factors: political, social, technological and cultural that may influence the sustainable development of bamboo industries in Ghana.

FACTORS THAT WILL INTERPLAY IN GHANA’S BAMBOO INDUSTRY

Undoubtedly, bamboo resources abound in Ghana’s forest cover. Like any natural resource, renewable and non renewable, the availability and ‘free’ access to the resource by interested parties and stakeholders may result in unsustainable exploitation. To ensure that such industry survive and provide the needed revenue for socio-economic well being of its citizenry a number of essential parameters are required for the sustainable development of bamboo industry. Among other things, the implementation of policies and management system of bamboo resources; adoption of a utilization tool; research and technological innovation and awareness creation have been highlighted as important factors that will interplay toward a viable bamboo industry especially in southern Ghana where the resource abound. Policies and management system of bamboo resources in Ghana. The systematic management of bamboo resources to ensure adequate and continuous supply of raw materials over a long period of time and the need for systematic exploitation to increase the production of bamboo stock has been highlighted by a number of researchers [4, 5]. In Ghana, the Ministry of Lands

and Forestry has embarked on a vigorous bamboo development programme (BARADEP) to increase biodiversity of bamboo species and diversify the use of bamboo species for various endproducts. Policy that enhances sustainable extraction and utilization of bamboo resources is needed and the periodic review of such policies and management system that will ensure equity in the sharing of profits that will accrue from the extraction and utilization of bamboo resources in different regions of Ghana. Micro-finance for entrepreneurs Capital investment and capacity building of local people interested in the establishment and management of bamboo factory is vital to the sustainable development of the industry especially in the developing country like Ghana. Currently the scheme of various financial institutions does not encompass bamboo processing and its industrial development. Considering the increasing trade values in bamboo products and the huge market for bamboo product in the era of green construction, conscious drive by entrepreneurs will demand financial institutions to help promote the establishment of such industry for job employment and enhanced livelihood of Ghanaians.

263


Research and Technological Innovations

Research and development have facilitated processing of bamboo in Asian countries from the simple use of raw bamboo culms to comprehensive exploitation of the resources for making handicrafts, ply bamboos, pulp and paper, drinks and medicines [6,7]. Considerable information has accumulated on various aspects of management and utilization of bamboo in Asia and this has facilitated processing of raw bamboo culms in Asian countries to ply bamboos, and pulp and paper. Governmental and non governmental organization like ITTO, International Development Research Center (IDRC) of Canada and International Network for Bamboo and Rattan (INBAR) have over the years supported several bamboo research and development projects mainly in Asia and in South America. The full utilization of bamboo in Ghana depends of the thorough knowledge of the technical properties of native bamboo species. Data and information on some essential technological properties of bamboo species in different ecological zones of southern Ghana have been studied. Specifically, the micro and ultra microstructure, physical, thermogravimetric behaviour, chemical and phytochemical properties of bamboo species were studied [8, 9, 10]. However, the machining and mechanical properties of bamboo (both natural and plantation grown ones) and the development of bamboo products for housing and construction need urgent research attention in Ghana like other African countries where the resource abound.

Development and adoption of a bamboo utilization tool (BuT) In an effort to address the under-utilization of bamboo resources in Ghana, simply knowing technical properties is insufficient for the creation of a viable bamboo-based industry. It is clear that a tool to help determine the most appropriate uses for bamboo resources (both in natural stands and plantations) in a specific area and at particular times is needed in African tropical countries in which bamboo is common. The Bamboo Utilization Tool (BuT), which has been created to bridge the knowledge gap on bamboo use [11] should be adopted. The Bamboo utilization Tool (BuT) integrates geographical, technological and socio-cultural information pertaining to the quality, quantity and consumption patterns of native bamboo resources in Ghana to determine which products should be developed in a particular area. It is hoped it will help underpin a process towards the sustainable use of bamboo resources.

Awareness creation Traditionally, local people around natural resources and users of renewable natural resources like bamboo have always been considered as important element in the process of technology diffusion and assimilation [12]. The need for their perception and the essentialist as well as determinist notions of users in ensuring wider utilization of resources and technological advancement is necessary. It will help organize potential users of bamboo products in Ghana. Earlier survey of some bamboo growing communities in Ghana revealed the categories of respondents in terms of the frequency of use of bamboo products. Majority of the respondents in the selected communities were annual users of bamboo products. Awareness among communities of industrial bamboo products was relatively low. Increase awareness of the potential of bamboo

264


resources for poverty reduction will be vital to the sustainable development of a profitable bamboo industry in Ghana. CCr1

CONCLUSION Continuous and uncontrolled exploitation of few ‘economic’ timber species is a threat to the African forest especially in the wake of climate change concerns. The development of underutilized and non-timber forest products such as bamboo could reduce the pressures on the traditional economic timber species. The development of sustainable bamboo industries in Ghana will involve commitment of many stakeholders: researchers, financial institutions, governmental commissions and the local and foreign investors. Further collaboration among bamboo-growing communities and various stakeholders in relevant governmental and non-governmental organizations is recommended to ensure the development of the bamboo industries in Ghana for enhanced livelihood of Ghanaians as observed in most Asian bamboo industries.

REFERENCES 1. Romualdo L. Sta. Ana (2006) Bamboo development in Asia. Abstracts and papers from INBAR international workshop on bamboo for sustainable livelihood, Wuyishan China. 2. Tekpetey, S., Frimpong-Mensah, K. and Darkwa, N. (2007) Thermogravemetric behaviour and selected physical properties of Bambusa vulgaris in Ghana. Journal of Bamboo and Rattan 6 (3,4): 199–204.

3. UNIDO (2001) The European flooring market and prospects for bamboo industrial products from Ghana. United Nations Industrial Development Organization, New York, USA. 4. Liese, W., (1985) Bamboo: biology, silvics, properties, utilization. Deutchsche Gesellschsft fur Technische Zusammenarbeit, Eschborn, Germany. 132pp 5. Omar Ali, M., (1981) Bamboo production and utilization: Research and development on the production and utilization of bamboo in Bangladesh. In T. Higuchi (Ed.). Bamboo production and utilization. Proceedings of XVII IUFRO Congress, Kyoto University, Kyoto, Japan. 6. Austin, R., Levy D. and Ueda L., (1983) Bamboo. John Weatherhill, Inc. New York and Tokyo

7. ITTO, (1997) Tropical Forest Update. Vol. 7, No. 4, Yokohama Japan

265


8. Ebanyenle, E. and Oteng-Amoako, A. (2007) Site differences in morphological and physical properties of Bambusa vulgaris grown in Ghana. Discovery and Innovation Vol 19 (Afronet Special Edition No 3), pp. 222–225.

9. Tekpetey S.L., N. A. Darkwa and K. Frimpong-Mensah (2008) Bambusa vulgaris in Ghana: Chemical composition and phytochemical properties for enhanced utilization Journal of Bamboo and Rattan Vol. 7 3&4 pp 243-249. 10. Tekpetey S.L., K. Frimpong Mensah and N.A Darkwa (2007) Thermogravimetric behaviour and physical properties of Bambusa vulgaris in Ghana. Journal of Bamboo and Rattan - JBR Vol 6 No 3&4 pp199-203. 11.

Tekpetey, S.L. (2009) The Development of a Bamboo Utilization Tool. ITTO Tropical Forest Update. Fellowship Report pg 18/4 pg 20-21

12. Oudshoorn, N. and T. Pinch (2003) How users matters: The Construction of users and Technology.Cambridge,MA; MIT Press.

266


Energy Consumption Structure of the Hungarian Wood Industry Varga, Mihály; Németh, Gábor; Kocsis, Zoltán; Bakki-Nagy, Imre University of West Hungary, Faculty of Wood Sciences, Institute of Machinery and Mechatronics, Sopron, HUNGARY; email: [email protected]

ABSTRACT The energy used in woodworking plants can be classified basically into heat energy and electrical energy. Based on our survey, in the 1980’s the ratio of the heat and electrical energy shifted from the ordinary 80%/20% to 50%/50% in several modern factories. Within the electrical energy consumption, the highest portion is used by machines (45%-60%), dust-chip suction systems (22%-28%), and in many cases the auxiliary equipment of kiln dryers (which may exceed 10%). Machines and infrastructure share the heat demand using 55%/45%. The energy consumption structure of companies with different profiles in the wood industry was investigated (companies from the timber industry and furniture industry, parquet and wooden house manufacturers, briquette and pellet manufacturers). It is clear that the structure of the energy consumption can be analyzed precisely only for individual companies since each company is different. The best examples are the diversity of suction systems and the air feedback (filtration heat loss, controlled or non-controlled ventilator engine rev). Another good example is the kiln drying. If a furniture manufacturer uses dried raw material, the value of this energy-segment will be 0%, while using a kiln dryer may result in an electrical energy consumption of 10% and a heat demand of 45% compared to the overall energy consumption of the company.

INTRODUCTION The development of wood processing technologies, their modernization, the increase of the level of mechanization, the presence expansion, the application of cutting edge technologies, the increase of automatization, the of quality environment protection and labor health obligations all contribute to the increase of energy use. Naturally the technical modernization partly reduces the specific energy consumption. In the wood processing factories, the energy used is basically heat and electrical energy.

267


The electricity users in the technologies are as follows: a) Woodworking machines and woodworking equipment, as well equipment required to maintain machines and equipment. b) Materials handling, storage of machinery and equipment. c) Extension of the technological process, auxiliary mechanical equipment, such as:  general building engineering  lighting  caloric devices  compressed air supply  hydraulic equipment  dust and chips extraction system  steam-gas extraction system  briquetting and pellet making machines  any other equipment (e.g. small machines) Heat users: a) Technology equipment  heat treatment equipment (dryers, steamers, wood modification etc.)  presses  surface treatment equipment b) Heating, hot water supply equipment

MEASUREMENT The presentation of the current structure of the wood industry energy consumption in Hungary First, the structure of energy consumption in Hungary is briefly the following. 70% of primary energy (1100 PJ) is the final energy consumption (794 PJ). The efficiency is about 44% from this 31% of useful energy (350 PJ). The final consumption is 18% of the industrial sector's energy consumption (142 PJ). On the basis of our measurements, we have defined the ratio of the total industrial energy use of half finished products in the sawmilling industry and their demand for energy. We started with the annual volume of the processed logs (Agricultural Office data: 1.6 million m3) The calculated value represents a very small proportion of the total industrial energy consumption (~1%, 1.33 PJ). Our research indicates that when the heat and electrical consumption is probably 50%-50%, the electricity consumption is approximately 105,000 tons of CO2 emissions. (Standard emission factor in Hungary: 0.566 tCO2/MWhe; source: “Technical supplement to the guide related to the planning for sustainable energy”). (If the current quota prices are taken into account that is more than 450 million Forints in the carbon market.) The heat energy consumption is approximately 37,000 tons of CO2 emissions if we use natural gas for heat production (standard emission factor in Hungary: 0.202 tCO2/MWhnatural gas). This means that 1 ton of wood processing creates about 150 kg CO2 emissions. Naturally if the heat production locally and that made from waste wood, then this value may be less due to the neutrality of the wood. (According to our estimates about seventy percent of the heat thus produced). The situation would be even more ideal if in the factories were using sun and geothermal energy. (We will develop implementation at the next phase of research). In part of this research, the various wood industry companies were analyzed (saw industrial, furniture industry, parquet manufacturer, wooden houses manufacturer as well as briquettes and pellet production companies). We have analyzed the company’s energy consumption. The

268


surveys have appeared markedly in those areas where the levels of energy consumption are large and the CO2 emission reduction can be done. Our investigations informed us about the energy use of wood industry companies, which are summarized as follows. Figures 1 and 2 show examples of the energy consumption in the saw and furniture industry. Naturally, on the figure the total installed electrical capacity isn’t seen, but the electrical capacity of those machines which are working at the same time. This value is about 45%-50% in small and middle sized wood industry companies in Hungary. In the wood industry, production equipment, uses 45%-60 % of electrical consumption; dust and chip extraction systems use 22%-28% and in some facilities the dryer’s fans use about 10%, (if the company doesn’t buy dry material). Dust and chip extraction and dryers use a significant share of the energy. Special attention must be given to these areas, and we propose a solution based on surveys, to reduce energy consumption. Extraction 22%

Compressors 7% Dryers and auxiliary equipment 2% Boiler auxiliary equipment 3%

Technology 60% Social and auxiliary equipment 6%

Figure 1 The energy distribution of the sawmills

Extraction 28%

Technology 45%

Compressors 6% Dryers and auxiliary equipment 11%

Boiler auxiliary equipment 4%

Social and auxiliary equipment 6%

Figure 2 The energy distribution of the furniture factories In our research, we emphasized the energy consumption of drying; therefore, we examine the ratio of electric and heat energy in the drying. Figure 3 shows the convection heat drying and electric power distribution.

269


Electrical energy demand 20%

Heat demand 80%

Figure 3 The convection heat drying and electric power distribution Based on our calculations, drying of wood per unit volume (using the same parameters) 20% of the minimum required invested energy is electrical energy and 80% of energy is heat energy in convection drying. We built an industrial test environment for the subsequent research that we will be able to use for energetic measurement. This system shows an exact picture from the real drying heat and electrical energy consumption for different wood species and moisture content.We summarized in a table the measured values from the initial test period. The calculated values, which are from the published formulas and from the surveys, approximately correspond to the real measured values. Table 1 Summary table of the drying energy demand

Denomination

Convection dryer (max. capacity 90 m3; drying time: 22 days) Vacuum dryer (max. capacity: 8 m3; drying time: 6 days)

Total energy demand The electrical The heat quantity /Oak/ energy rate Initial moisture content: 45% rate from the from the total Moisture content after drying: total energy energy necessity 8% necessity [%] [%] [MJ/m3]

Theoretical calculated energy demand /Oak/ [MJ/m3]

2377

12

88

2200

1248

40

60

Not found exact calculation method

The heat demand is shared between production and infrastructure equipment (55%-45%) Heating, hot water generating equipment heat demand: 54%

Technological heat demand: 46%

Figure 4 The heat energy demand distribution of a general furniture factory The surveys showed that we can precisely analyze the structure of energy consumption only at the particular companies, because every company is different. The best examples are different extraction systems and specific solutions of energy recovery. (We consider only the filtration heat loss but from a more complex examination, it matters if a fan motor speed is regulated or not). Another example is the drying. If a furniture industry buys dried wood then “the slice of energy cake” is 0%, but if the company dries wood, then 10% of total energy is electrical and 45% of

270


total energy is heat energy. The measurement of exact quantity of heat used can cause many problems because many companies generate heat from wood waste (locally generated) and this isn’t registered anywhere. (The energy consumption can be inferred from the wood waste used as fuel, performance of the boiler and from the working time). Electric energy required for a unit of base material Figure 5 shows how much electric energy is needed to produce a unit of base material/product and how it differs between companies. We have defined the given values mathematically in the function of the quantity of the base material and the finished product. To produce one cubic meter of product, more energy is needed than to process one cubic meter of base material. This means the change of production costs among the different company groups. The differences between the production bar, and the base material bar, come from the different amount of waste material and secondary products. Closer inspection shows that the ratio of the use value impacts the energy costs used in manufacturing a product. In furniture production the energy use of the sawmilling industry needs to be included since most furniture manufacturers work with sawn base material. This is generally bought in a dry condition. Small and medium enterprises buy sawn lumber as undried in order to increase their cost effectiveness. They dry the products later themselves. Distribution of electrical energy (unit of base material and unit of product)

1200 )h 1000 W k (y 800 g re n el ac 600 rit ce l E 400

The energy required to processing 1 m3 of base material

The energy required to preparing to 1 m3 of product

200 0 Furniture industry

Wooden house manufacturer

Parquet manufacturer Pellet manufacturer

Sawmill

Figure 5 Data from woodworking companies about their distribution of the energy demand which is needed to produce to unit of base material or unit of product (companies which included in our survey) Main factors for the exceptionally high energy consumption in the furniture industry:  Processing unit of base material and product development is much more time consuming and also needed more equipment.  Generally better implemented all of the dust and chips extraction systems (usually each machine has its own extractor) and therefore, needs more electrical energy. The same is to be treated in the sawmill industry.  Today, the SME furniture companies have better infrastructure, which had an impact on energy consumption.

271


Industrial test environment We built an industrial test environment for the subsequent research. It has been integrated into heat, electrical energy meters and analysis system (Figure 6). We surveyed the energy structure of a typical Hungarian company. Then we can offer effective solutions to reduce energy consumption. The solutions to reduce energy consumption will be easy to adopt for the various companies during the project, whether large companies or SME.

Figure 6 Industrial test environment The measurement system has been built for measuring quantities of electricity and heat. A measuring device has been installed on the energy consumer which we want to measure. These devices communicate through a bus system with a bus controller which transmitted the measured data to a computer. Software saves the data which are necessary. The data can be seen on the measuring device and on the computer. The saved data can be queried by several criteria.

Figure 7 A schematic diagram of measuring system

272


When we were done with installation, we started the testing (with 8 pieces electrical and 2 heat measuring devices). Our current financial situation did not allow us to buy more instruments. Therefore, only the dominant energy consumers were selected. When there were more of the same type energy consumers, we chose one of them and we examined only that. In the next part we will show the electric and heat energy use of certain consumers with the help of data from a measuring system. Since we have set the sampling intervals at 5 and 15 minutes we got a data line of 3000-9000 lines. These data lines show year, month, day, hour, minute and seconds. These can easily be adjusted to the technology. To simplify the evaluation, we show the amount of electricity and heat used by machines and equipment over monthly periods. Since we do not collect data from each machine, we did the missing measurements ourselves with a hand operated benchmarking device. During testing we created a map of one month’s energy use and this is shown as follows:

Figure 8 The map of test environment

CONCLUSIONS We concluded that we can achieve significant decreases in energy consumption. These include:  Dryers: Our preliminary surveys show that these devices have the largest energy consumption (electrical and heat energy), so substantial energy savings can be made. We can use alternative energy sources (reducing energy consumption about 30% in summer) and observe the drying schedules. Our research reviewed the dryer’s external structure too, because a dryer with smaller overall heat transfer coefficient needed less heat. (The transmission heat loses are high now). We will do thermo vision tests in future research.  Dust and chips extraction systems: Heat and electrical energy can be more efficiently used in the wood industry’s dust and chips extraction systems. Here, the electrical energy is consumed by fans, separators and the operation of other ancillary equipment. The heat

273


energy is consumed by filtration losses. (Usually the filtration heat losses are more than the transmission heat losses). For these reasons, we started comparing energy consumption in the operation of the traditional system used by 90% of the factories and a flexible technology solution.  Space Heating: Heat loss: heat flux from across the room’s wall. This amount is high in the current wood industry, because the structure of the walls, windows and doors are obsolete. The proper insulation would reduce heat transmission approximately by 15%20%.  Compressors: The recovery of generated heat for space heating during their operation.  Ways to reduce electricity consumption in production lines: Production lines use more than half of the factories’ energy. This can be reduced by rationally coordinating the processing and reduce the machines and grouping of machines simultaneity. Another substantial energy reduction options is the examination of the processing machines’ motors. Generally motors are built into the machines that are more powerful than needed. They need more electricity, which deteriorates the efficiency of energy use increasing the energy cost. One solution is newer motors with a better power factor. The control of frequently used machines will reduce electricity consumption by 30%.  Use of the electronic energy management system and power operator system: The system turns equipment turning on and off keeping consumption within fixed limits instead of instantaneous power consumption. Therefore, the electricity supplier does not need such a large fixed capacity. This avoids the substantial penalties for exceeding limits of consumption. For example the processing control systems of the dust and chip extraction system are easily amenable to the management operation of the system. The Institute of Machinery and Mechatronics has developed an innovative monitoring system, which increases factory safety and reduces power consumption. The energy monitoring falls within the working area, buildings, lighting regulation (to optimize the light in the breaks or turning off unused lighting in workshops; insulating technical equipment).

REFERENCES 1. Mezőgazdasági Szakigazgatási Hivatal Erdészeti Igazgatóság (2010) Erdővagyon, erdőés fagazdálkodás Magyarországon, Leporelló, Budapest. 2. Petri L. (2003) Energiatakarékos fűrészáru szárítás, Szerzői Kiadás, Budapest.

274


A Novel Approach for Log Sawing Optimization Cao Pingxiang; Wang Yi; Huang Fei; Guo Xiaolei College of wood science and technology, Nanjing Forestry University, Nanjing,210037, P.R.China Abstract: The lumber industry has long been suffering from wood supply problems, including source shortage and poor quality. One of the important ways to alleviate the problem is to increase lumber-production rate by utilizing the well-developed optimization theory. This paper presents a new optimization approach for lumber manufacturing. The approach is developed through an analytical method to minimize the waste by a given objective function. The proposed optimization algorithm is based on the ladder cutting method. The application of the optimization approach is illustrated by an example. Keywords: Double circular saw; Quarter-sawn board sawing; Mathematical model; Optimal sawing 1. Introduction The cross sectional shapes of logs are influenced by log species and growth environment etc, but most section shapes are nearly circular. In order to improve timber-produced rate and plank grate rate, optimal sawing method is required. According to the specification and grade of lumber, the calculation of optimal sawing is based on the size and shape of logs small head end in an effort to get more high-grade dimension boards [1]. Optimization approaches have been widely utilized to maximize the output rate in manufacture of steel plate, glass, wood-based panel and logs sawing [2]. The optimal lumber sawing method is the most complicated problem which is nondeterministic polynomial (NP) completed, it is difficult to get the optimal solution within acceptable time. In recent years, approximate algorithm was often used in optimal layout of logs sawing, which had reached higher stability, universality and efficiency etc while obtaining an optimal material utilization, but the existing approximate algorithm was often less than satisfactory, especially to obtain the maximum yield rate of quarter-sawn board. To overcome these limitations stated above, a new optimization algorithm for log sawing using double circular saw is presented in this paper. Assuming circular cross-section of log small head and quarter-sawn board as product, the algorithm uses the optimal layout method on plane to obtain maximum quarter-sawn boards and minimum surplus materials. 2. Double circular sawing Double circular saw is a numerically controlled log sawing equipment that is based on the ladder cutting method and adopts optimal layout algorithm of quarter-sawn board to realize the optimal sawing. It has horizontal and vertical sawing frames and the two circular saws are configured at a right angles. There is some distance between the two saws along the axis of log, and the projection of their cutting edges intersect at one point on the log cross-section. The two saws will be moved to calculated positions along the x- and y- directions simultaneously, With the double circular saw moving along the z-direction, a piece of plank was sawed away. Based on the optimal ladder cutting method, the horizontal and vertical saws are moved from left to right and top to down respectively to obtain quarter-sawn boards with precision. The cutting principle is illustrated in Figure 1.

275


Fig.1 The cutting principle of double circular sawing

3. Mathematical model of the optimal layout of quarter-sawn board The optimal layout of quarter-sawn board is the foundation of double circular saw under the numerical control NC. Its optimal layout can be described as follows: R is the radius of the inscribed circle of the log small head end . Let r denotes the radius of the maximum inscribed circle centered at the pith of small head end. Let

L denote the length of saw log. Let W denote the saw kerf, which is the space between adjacent planks. In order to produce H × W size dimension boards ( H and W are the width and thickness of dimension board respectively), after the optimal layout of dimension boards has been obtained, whether nesting 1/2、1/4 size dimension boards (1/2 H × W and 1/4 H × W dimension boards) can be decided to improve the timber-produced rate, The objective of optimization is to get the maximum yield rate of quarter-sawn board and the minimum surplus materials. The objective function of optimal layout, or the minimum surplus materials function is

Sj 

 R 2  ( pWH  1/ 2  qWH  1/ 4  vWH )  J  K  R2

where p 、q and v respectively represent the numbers of dimension board、1/2 and1/4 size dimension boards on small head end. The total saw kerf loss is J  p (2 H  2W ) w  q ( H  2W ) w  v(1/ 2  H  2W ) w . For logs with large pith, the pith will be excluded as waste with the loss of K  r The sawing problem can be modeled with analytical method of mathematical programming, with ordered number pairs corresponding to points in plane or space in rectangular coordinate system. Then algebraic method could be used to study the shape, size and position of plane or space graphic to develop the mathematical model of logs optimal sawing[3]. 2

Select the centre of a circle of log as layout coordinate origin (0, 0). The pith eccentricity was ( x p , y p ). Dimension boards were arranged within the inscribed circle of log small head end. The position of dimension boards can be determined by the lower left point ( xi1 , yi1 ) and the upper right point ( xi 2 , yi 2 ) of the rectangular planks. In this model quarter-sawn board is defined as the plank sawed along the radial direction of log, and the angle between tangent of annual ring and broadside is more than 45 º .Two cases existed: 1）The pith had no eccentricity , namely ( x p , y p ) = (0, 0) . In this case the annual ring was relatively regular,

276

20th International Wood Machining Seminar quarter-sawn board could be defined by regarding the annual ring as perfectly circular. With two lines: y  x and y  - x , log small head end was divided into four areas. As shown in Figure 1, lumbers cut horizontally in area A and C and lumbers cut vertically in area B and D are quarter-sawn boards. (Planks layout could only be horizontal or vertical because of the sawing process of double circular saw, for vertical layout, the long side of the r plank is parallel to Y axis and for horizontal layout the short side is parallel to Y axis); 2）The pith had eccentricity, namely （ x p ， y p ）=（a，b）. In this case, the annual rings are irregular and concentrate in the eccentricity side with a and b not equal to zero at the same time. Quarter-sawn board could be defined by regarding the pith as center as follows : Log small head end is divided into four areas by two lines y - b  x - a and y - b  - x  a , which pass through pith. As shown in Figure 2, lumbers cut the horizontally in area A’ and C’ and cut vertically in area B’ and D’ are quarter-sawn boards.

Fig.2 The definition of radial board

The layout is confined by the quarter-sawn board definition, and the ideal situation is that layout in every demarcation area can be configured according to the definition of quarter-sawn board. Let a point on the nesting of quarter-sawn board be（x，y）,so for the layout in each area, the ideal nesting of dimension boards that meets the definition of quarter-sawn board could be modeled for two circumstances: the pith had no eccentricity, and the pith had eccentricity : Nesting of dimension boards when the pith had no eccentricity: Area A: S . t. x  y, and x  - y , xi 2  xi1  W , yi 2  yi1  H ； Area B: S . t. x  y, and x  - y , xi 2  xi1  H , yi 2  yi1  W ； Area C: S . t. x  y, and x  - y , xi 2  xi1  W , yi 2  yi1  H ； Area D: S . t. x  y, and x  - y ； xi 2  xi1  H , yi 2  yi1  W 。 Nesting of dimension boards when the pith had eccentricity: Area A : S . t. x - a  y - b, and x - a  -( y - b) , xi 2  xi1  W , yi 2  yi1  H ； Area B : S . t. x - a  y - b, and x - a  -( y - b) , xi 2  xi1  H , yi 2  yi1  W ； Area C : S . t. x - a  y - b, and x - a  -( y - b) , xi 2  xi1  W , yi 2  yi1  H ； Area D : S . t. x - a  y - b, and x - a  -( y - b) , xi 2  xi1  H , yi 2  yi1  W 。

277

20th International Wood Machining Seminar Nesting of 1/2 size dimension boards when the pith had no eccentricity: Area A: S . t. x  y, and x  - y , xi 2  xi1  W , yi 2  yi1  1/ 2  H ; Area B: S . t. x  y, and x  - y , xi 2  xi1  1/ 2  H , yi 2  yi1  W ; Area C: S . t. x  y, and x  - y , xi 2  xi1  W , yi 2  yi1  1/ 2  H ; Area D: S . t. x  y, and x  - y , xi 2  xi1  1/ 2  H , yi 2  yi1  W . Nesting of 1/2 size dimension boards when the pith had eccentricity: Area A : S . t. x - a  y - b, and x - a  -( y - b) , xi 2  xi1  W , yi 2  yi1  1/ 2  H ; Area B : S . t. x - a  y - b, and x - a  -( y - b) , xi 2  xi1  1/ 2  H , yi 2  yi1  W ; Area C : S . t. x - a  y - b, and x - a  -( y - b) , xi 2  xi1  W , yi 2  yi1  1/ 2  H ; Area D : S . t. x - a  y - b, and x - a  -( y - b) , xi 2  xi1  1/ 2  H , yi 2  yi1  W . Let the coordinates of the lower left point and the upper right point of the any two rectangular planks — a and b be ( xa1 ，ya1 ), ( xa2 ，ya 2 ) and ( xb1 ，yb1 ), ( xb 2 ，yb 2 ) respectively. The rectangular planks will not overlap under following conditions:

xb1  xa 2 ；or xb 2  xa1 ；or yb1  ya 2 ；or yb 2  ya1 。 For any plank to lie within the log, following conditions must be met:

x  R, y  R, a 2  b 2  R 2 , xi12  yi12  R 2 , xi 2 2  yi 2 2  R 2 。 The nesting of 1/4 size dimension boards is the same as the standard quarter-sawn boards. 4 The principle of searching optimal layout of quarter-sawn board The searching of optimal layout of quarter-sawn board is the process to realize the algorithm of optimal layout. An optimization algorithm is actually a search process or rule based on some ideas and mechanisms to get solution to meet the users’ requirements through certain ways or rules [4]. The mathematical model of the optimal layout of quarter-sawn board described the mathematical realization mechanism of the nesting of quarter-sawn board in ideal situation, namely nesting in every demarcation area according to the definition of quarter-sawn board. Because the layout can’t go beyond the border in the practical process, the materials in boundary area will be wasted. In order to solve this problem, semi quarter-sawn board can be configured in the area near the border after the nesting of quarter-sawn board when initial layout area is determined. Through the local optimal solution to search for the optimal layout, the maximum yield rate of quarter-sawn board can be ensured. In other words, the layout way of semi quarter-sawn board in boundary area is changed to calculate the ratio of quarter sawing component in the plank (the proportion of the quarter sawing in a plank), the highest proportion of the quarter sawn parts will be the optimal layout in the boundary area. The computation effort will be tremendous due to the large numbers of nesting of planks. In order to accelerate searching speed, some constraints are set to simplify the search process: the nesting of dimension boards is preferred, after which the nesting of 1/2 dimension boards and 1/4 dimension boards will be configured when there is not enough space; The layout space between adjacent planks is w, the planks cannot overlap and

278

20th International Wood Machining Seminar exceed the boundary;

The pith is set as the starting point of the layout.

The nesting of quarter-sawn board is

preferred in area D or D and the other dimension boards in blank areas will be configured after the nesting of quarter-sawn board had finished. The horizontal or vertical layout will be determined based on local optimal solution to ensure the nesting of quarter-sawn board. 5. Computer aided nesting Based on the graphics operation and computational geometry, the logs and lumbers are processed as graphics .It used the knowledge of graphics and computational geometry to position and deal with the layout materials; Combined with the layout algorithm to search the optimal layout [6], the optimal layout is an intelligent way to realize the optimal sawing with double circular saw. The computer aided nesting with double circular saw realized the optimal layout of quarter-sawn board by programming. The specification of planks can be stored in computer. When the outline parameters of logs are input to the computer, it will automatically analyze the specification of planks and optimize the sawing plan. Then lumbers are sawn under numerical control according to the optimal sawing plan. The computer aided nesting system of double circular saw mainly includes the following: (1) The layout information processing: read the log information, including diameter, length, diameter of pith, pith eccentricity; read the plank information including width, thickness, and read auxiliary information including whether arranging 1/2 or1/4 dimension boards, and saw kerf, etc. (2) The optimization calculation: collect the information of the nesting of logs and dimension boards as the initial input information for generating all optimization algorithm modules. Start the optimization calculation and collect all optimization methods to get the best result to form final result data of optimal layout. (3) Generating the layout diagram: the data was extracted from some good results to generate layout diagram. (4) Exporting the layout result: export the layout diagram and the statistical information of material consumption; Return the layout result information to the production operation management system; program NC code for each plank according to the operation instruction. The optimal layout plan of all kinds of dimension boards could be obtained with computer aided nesting by programming. Based on the optimization algorithm of quarter-sawn board, some examples of the nesting of quarter-sawn board are shown in Fig.3. The diameter of the log is 500 mm , length is 1500 mm , width and thickness of plank are H = 20 mm, and W=60 mm respectively, the saw kerf is 2.5 mm, Figure 3 (a) and (b) show the radius of the pith is 0mm, figure 3 (a) is the case that the pith has no eccentricity and Figure 3 (b) shows the pitch has eccentricity (a=30mm, b=50mm); Figure 3 (c) and (d) show the result for logs with pith diameter of 40 mm, figure 3 (c) is the case for the pith have no eccentricity , and figure 3 (d) is for the pith have eccentricity (a=30 mm, b=50 mm).

（a）the radius of the pith is 0mm and have no eccentricity

（b）the radius of the pith is 0mm and have eccentricity（50，30）

279


（c）the radius of the pith is 40mm and have no eccentricity （d）the radius of the pith is 40mm and have eccentricity（50，30）

quarter-sawn board

semi quarter-sawn board

flat-sawn board

Fig.3 The nesting results of the log with diameter 500mm with CAN

The results of computer aided nesting show that: the larger the diameter of logs and the smaller specification of planks in a certain range, the higher the timber-produced rate that can be obtained. But when the specification of planks is smaller, the saw kerf loss will increase, and the timber-produced rate will also reduce; When the length of planks is longer, the timber-produced rate will be lower, but the lumber recovery will be significantly increased by increasing the nesting of 1/2, 1/4 dimension boards. When the length of logs is less than 2 m , the timber-produced rate from optimal sawing pattern can reach more than 60% ,which is higher than the average level in China —55% [7] , Therefore, the optimization greatly improved the utilization of logs, ensured the yield rate of quarter-sawn board, and increased the value of sawn timber product. 6. Conclusion This paper presents a novel optimization algorithm for log sawing problem with double circular saw. A mathematical model was developed to describe the problem. Then, the mathematical realization mechanism of optimal layout is described. The optimal layout of sawing logs is implemented through computer-aided nesting. The experiment result shows that higher yield rate of quarter-swan board can be achieved by applying optimization theory to the sawing technology. What’s more important, this approach is promising to provide the theoretical foundation for more intelligent NC sawing equipment.

References [1] Xiong Ying, Rao Singiresu S.. A fuzzy dynamic programming approach for the mixed-discrete optimization of mechanical systems[J]. Journal of Mechanical Design, 2005, 127(6):1088-1099. [2] Zhao Hui, Xi Ping. The study of the rectangular optimal layout algorithm and the system [J]. Forging technology, 2005,47(1):19-22. [3] Yuan Yaxiang, Sun Wenyu. The optimization theory and method [M].Beijing: Science press,1997:100-102. [4] Cao Ju, Zhou Ji. An approximation algorithm of the rectangular optimal layout [J]. Computer aided design and graphics, 1995,7(3):190-195. [5] Ann Van DerWilt. An Algorithm for Two Stage Unconstrained Guillotine Cutting[J]. European Journal of Operational Research, 1995, 84(2): 494-498. [6] Cui Yaodong. Computer nesting technology and application [M]. Beijing: Mechanic industry Press,2003:20-40. [7] Luo Xichun, Ma Yan. Development Trend of Foreign Countries Timbermaking Industries and Preliminary Study of Development Model of Chinese Timbermaking Industries [J]. Forestry Machinery & Woodworking Equipment,1997,25(5):4-9.

280

8. Chip Formation Oral Presentations


Application Of Fracture Mechanics For Energetic Effects Predictions While Wood Sawing Orlowski, Kazimierz1( ), Ochrymiuk, Tomasz2, and Atkins, Anthony3 1

Gdansk University of Technology, Faculty of Mechanical Engineering, Department of Manufacturing Engineering and Automation, Gdansk, Poland, [email protected] ( ) 2 The Szewalski Institute of Fluid-Flow Machinery, Polish Academy of Sciences, Department of Transonic Flows and Numerical Methods, Gdansk, Poland, [email protected] 3 University of Reading, School of Construction Management and Engineering, Whiteknights, Reading, Berkshire, United Kingdom, [email protected]

ABSTRACT In the classical approach, energetic effects (cutting forces and cutting power) of wood sawing process are generally calculated on the basis of the specific cutting resistance, which is in the case of wood cutting the function of more or less important factors. On the other hand, cutting forces (power) could be considered from a point of view of modern fracture mechanics. Cutting forces may be employed to determine not only toughness but also shear yield strength, which are then applied in the models. Furthermore, forecasting of the shear plane angle for the cutting models, which include fracture toughness in addition to plasticity and friction, broaden possibilities of energetic effects modelling of the sawing process even for small values of the uncut chip. Mentioned models are useful for estimation of energetic effects of sawing of every kinematics. However, for band saws and circular sawing machines the chip acceleration power variation as a function of mass flow and tool velocity ought to be included in analysis of sawing at larger cutting speeds.

INTRODUCTION Theoretical and experimental determination of values of forces acting in the cutting process belongs to the basic and simultaneously the most developed field of mechanics of this process. A great number of theoretical works, which were improved and experimentally verified, have been devoted to this problem. In the classical approach energetic effects (cutting forces and cutting power) of wood sawing process are generally calculated on the basis of the specific cutting resistance kc (cutting force per unit area of cut) [1, 2, 3], which in the case of wood cutting is the function of the following factors: wood species, cutting direction angle (cutting edge position in relation to wood grains), moisture content, wood temperature, tooth geometry, tooth dullness, chip thickness and some others which are less important [4, 5]. Many of those traditional models are empirical and based upon limited information employing blades having standard thickness kerfs. Moreover, for each type of sawing kinematics different values of specific cutting resistance kc have to be applied [5]. On the other hand, cutting forces (power) could be considered from a point of view of modern fracture mechanics [6]. Atkins’s ideas [6] can be also applied in analyses of sawing processes in which the offcut formation by shear occurs, e.g. in the real sawing process on a sash gang saw with the method based on the macro-mechanics [7, 8, 9]. In this paper we are going to prove that cutting power models which are based on modern fracture mechanics are useful for estimation of energetic effects of sawing of every kinematics.

281


THEORETICAL BACKGROUND Making an assumption that cutting force Fc acting in the middle of the cutting edge is an equilibrium of forces related to the direction of primary motion for a single saw tooth the mechanical process of material separation from the sawn workpiece, i.e. chip formation, can be described by the example of an orthogonal process (two dimensional deformation). The forces acting on the tooth can be represented in the classical approach by Ernst and Merchant’s force circle shown in Figure 1. α

γ

Fµ

βµ

FNΦ

βµ− γf

FTΦ

Φc FN

Fc

f z (h)

Fa Ff

Figure 1: Simplified cutting process model with Ernst and Merchant’s force circle [7]: Fa – active force, Fc – cutting force, Ff – thrust force (passive), Fµ – friction force on the rake face, FN – normal force to the rake face, FTΦ Φ – the force required to shear the wood along the shear plane, FNΦ Φ – normal force on the shear plane, αf – clearance angle, Φ c - shear angle, γf – rake angle, βµ – friction angle According to Atkins [6] and Orlowski [8], furthermore, taking into account that the chips have to be accelerated to the same velocity as the cutting tool velocity vc [6, 9], cutting power for one saw blade during the cutting stroke on a sash gang saw, and during cutting on a bandsaw machine, because their sawing kinematics are similar (Figure 2a), has the following mathematical formula:

 H  τ Sγ  H  RSt  P cw = Fc vc + Pac = Ent  P  ⋅ γ t vc f z + Ent  P  ⋅ vc  + Pac  P  Qshear    P  Qshear

(1)

H  where: Ent  P  is a number of teeth being in the contact with the kerf (integral), Hp is  P  workpiece height (cutting depth), τγ is the shear yield stress, γ is the shear strain along the shear plane, which is given by:

γ=

cos γ f

cos(Φ c − γ f )sin Φ c

(2)

fz is feed per tooth (uncut chip thickness h), St is a kerf (the width of orthogonal cut), βµ – friction angle which is given by tan-1µ = βµ, with µ the coefficient of friction, γf is the rake angle, Φc is the shear angle which defines the orientation of the shear plane with respect to cut surface, R is specific work of surface separation/formation (fracture toughness), and Qshear is the friction correction:

282


Qshear = [1 − (sin β µ sin Φ c / cos( β − γ f ) cos(Φ c − γ f ))]

(3)

For least force Fc the shear angle Φc satisfies [10]:

   sin β µ sin Φ c 1 1  − 2 1 − ⋅ 2 =  cos( β µ − γ f ) ⋅ cos(Φ c − γ f )   cos (Φ c − γ f ) sin Φ c 

(4)

 sin β µ  cos Φ c sin Φ c sin(Φ c − γ f )  = − cot Φ c + tan(Φ c − γ f ) + Z ⋅  +   cos 2 (Φ c − γ f )   cos( β µ − γ f )  cos(Φ c − γ f )

[

]

R is the parameter which makes Φc material dependent. Equation (4) is solved τγ ⋅ fz numerically [11]. in which Z =

The chip acceleration power Pac variation as a function of mass flow and tool velocity is given by: •

Pac = m vc2

(5)

•

where: m (kgs-1) represents the mass of wood (chips) evacuated in a certain period of time at the certain cutting tool velocity vc (cutting speed), which can be calculated as follows: •

m = H P St v f ρ

(6)

In Eq. (6) vf is feed speed and ρ is density of sawn wood. It should be emphasized, that in these analyses it was assumed that the power Pac is not a function of the number of working teeth. a)

b)

fz f

h

P

1

D/2

St

s

2

Figure 2: Sawing kinematics on the sash gang saw and band sawing machine (a), along with kinematics on circular sawing machine (b): fz – feed per tooth, s – saw blade thickness, AD – area of the cut, P – pitch, Y, Z and YM, ZM – machine coordinate and setting axes, Yf – f-set coordinate axis, Pfe – working plane, D – circular saw blade diameter, h – uncut chip thickness, Hp – workpiece height (depth of cut), a – position of the workpiece, ϕ – angular tooth position

283


Kinematics of sawing on circular sawing machines (Figure 2b) differs from kinematics of cutting on sash gang saws and bandsawing machines. In case of cutting with circular saw blades uncut chip thickness h (an average value e.g.) instead of feed per tooth fz should be taken into account, hence, the cutting power may be expressed as:   ϕ −ϕ  τ S γ  ϕ − ϕ  RSt  P cw = Fc vc + Pac = Ent  2 1  ⋅ γ t vc h + Ent  2 1  ⋅ vc  + Pac ϕ Q ϕ Q t shear t shear      

(7)

 ϕ −ϕ  where: Ent  2 1  is a number of teeth being in the contact with the kerf (integral), ϕ1 is an  ϕt  2(H p + a ) , ϕ 2 is an exit angle which can be angle of teeth entrance which is given by ϕ1 = arccos Dcs 2a determined as ϕ 2 = arccos , Dcs is a diameter of circular saw blade, an average uncut chip Dcs thickness is given by h = f z sin ϕ , and an average angle of tooth contact with a workpiece ϕ is ϕ + ϕ2 calculated from ϕ = 1 . 2 Table 1: Tool and machine tool data Parameter

Sash gang saw HDN (f. EWD)

Circular sawing machine HVS R200 (f. HewSaw)

Bandsawing machine EB 1800 (f. EWD)

Hp [mm]

140

80

140

St [mm]

4.0

3.6

3.1

P [mm]

25

–

50

γf [º]

14

22

28

z [–]

32

24

217

vc [m/s]

6.4

64.14

35

0–15 (0–0.25)

70–150 (1.16–2.5)

40–70 (0.67–1.167)

fz [mm]*

0–1.953

1.23–1.56

0.95–1.67

h [mm]*

0–1.953

0.59–0.747

0.95–1.67

PEM [kW]

160

2×90

4×110

vf [m/min] ([m/s])

*The values used in computation of predicted cutting powers

MATERIAL AND EXAMINED SAWING MACHINES DATA Predictions of cutting powers have been made for the case of sawing on three types of basic sawing machines such as the sash gang saw (HDN, f. EWD), the band sawing machine (EB 1800, f. EWD) and the circular sawing machine (HVS R200, f. HewSaw), which are used in Polish sawmills. The basic sawing machines data and cutting parameters for which computations were

284


done are shown in Table 1. Computations were carried out in each case study for one saw blade. The raw material was pine wood (Pinus sylvestris L.) of depth of cut equal to Hp (Table 1) derived from the Baltic Natural Forest Region in Poland. The raw material indispensable data for computation such as: fracture toughness R = 840 Jm-2 and the shear yield stress τγ = 22636 kPa was determined according to the methodology described in the works [12, 13]. The latter tests of raw material data determination were carried out on the sash gang saw PRW15M [14 ]with satellite tipped saw blades with a kerf equal to St = 2 mm. The average density of samples was ρ = 525 kgm-3, at moisture content MC 8.5–12%. A value of friction coefficient µ = 0.6 for dry pine wood was taken from the work [15]. The chip acceleration power Pac variation was estimated for each kind of sawing kinematics for depth of cut equal to Hp = 100 mm.

RESULTS AND DISCUSSION Predictions of cutting model that includes work of separation in addition to plasticity and friction in the case of sawing dry pine wood on examined sawing machines are shown in Figure 3. The reductions in Φc (Figure 3a) and increases in γ (Figure 3b) are visible in plots, concerns what happens near the origin of both Φc vs. h ( fz) and γ vs. h ( fz) plots. Those changes at small depths of cut are the reasons for the increase in cutting pressure for small values of feed per tooth (the so-called ‘size effect’ ) [6, 10, 15]. Furthermore, an increase in shear plane angle Φc is observed when rake angle γf has a larger value. b) 50

10

45

9

40

8

shear strain along the shear plane γ

shear plane angle Φc [deg]

a)

35 30 25 gamma 9 gamma 15 gamma 22 gamma 28

20 15 10 5 0 0.00

7 6 5

gamma 15 gamma 22

4

gamma 28

3 2 1

0.20

0.40

0.60

0.80

1.00

1.20

0 0.00

0.20

0.40

0.60

0.80

1.00

1.20

uncut chip thickness [mm]


Figure 3: Predictions of cutting model that includes work of separation in addition to plasticity and friction in the case of sawing dry pine wood on examined sawing machines (a) shear plane angle Φc vs. fz, (b) primary shear strain γ vs. fz, where: gamma 9 (tests on sash gang saw PRW15M), gamma 15 (sash gang saw HDN), gamma 22 (circular sawing machine HSV R200), gamma 28 (bandsawing machine EB1800) are rake angles In Figure 4, the chip acceleration power Pac variations as a function of feed speed vf and cutting speed vc for the sash gang saw HDN (Figure 4a), the circular sawing machine HSV R200 (Figure 4b) and the bandsawing machine EB1800 (Figure 4c) for the cutting processes with one saw blade while sawing dry pine wood for depth of cut equal to Hp = 100 mm are presented. For sash gang saws a maximum value of the chip acceleration power Pac equals to ≅ 2.5 W. Thus, in those machine tools where cutting speeds and feed speeds are rather small if compared with circular saws and band saws, chip momentum may be disregarded, the same as in the case of metal

285


cutting where it is customarily ignored [6]. In case of both the circular sawing machine and the bandsawing machine the chip acceleration power Pac is several hundred larger in comparison with the sash gang saw. a)

b)

c)

Figure 4: Predictions of chip acceleration power variation Pac as a function of cutting speed vc and feed speed vf for sawing of the pine workpiece of 100 mm in height with one saw blade on sash gang saw HDN (a), circular sawing machine HSV R200 (b), and band sawing machine EB 1800 (c) Comparison of predictions of cutting powers obtained with the use of cutting models that include work of separation in addition to plasticity and friction, and chip acceleration power variation in the case of dry pine sawing with one saw blade for three examined typical sawing machines are shown in Figure 5. Obtained values for those machines seem to be reasonable if compared to the power PEM of installed electric motors (Table 1), and also they are in conformity with values calculated with the use of empirical calculation models. Furthermore, they proved that predictions of cutting powers obtained with the use of cutting models that include work of separation in addition to plasticity and friction together with the chip acceleration power variation is a useful tool for estimation of energetic effects of sawing of every kinematics.

286


CONCLUSIONS The conducted analyses of energetic effects with the use of cutting models that include work of separation in addition to plasticity and friction corroborated their versatility and revealed the usefulness for every known type of sawing kinematics. Moreover, in the estimation of cutting power for sash gang saws chip momentum may be disregarded. On the other hand, in case of cutting on band sawing machines and circular sawing machines the chip acceleration power Pac has to be taken into account. 30000

Circular sawing machine HSV R200

25000

cutting power [W]

20000

15000

Bandsawing machine EB 1800 10000

5000

Sash gang saw HDN 0 0.00

0.50

1.00

1.50

2.00

2.50


Figure 5: Comparison of predictions of cutting powers obtained with the use of cutting models that include work of separation in addition to plasticity and friction (lower lines), with chip acceleration power variation added (upper lines) in the case of dry pine sawing with one saw blade

ACKNOWLEDGEMENT The author1 would like to acknowledge firms PPH GASSTECH Sp. z o.o. (PLC, Suwalki, PL) and Wydawnictwo Inwestor Sp. z o.o. (publisher, PLC, Tczew, PL) for their financial support for his participation at the 20th International Wood Machining Seminar.

REFERENCES 1. FISCHER, R. (2004) Micro processes at cutting edge – some basics of machining wood. pp. 191–202. In: Proceedings of the 2nd International Symposium on Wood Machining, Vienna, Austria. 2. SCHOLZ, F., DUSS, R., HASSLINGER, R., RATNASINGAM, J. (2009) Integrated model for the prediction of cutting forces. pp. 183–190. In: Handong Zhou, Nanfeng Zhu, Tao Ding

287


(Eds.) Proc. of 19th International Wood Machining Seminar., October 21–23, Nanjing, China, Nanjing Forestry University. 3. ORLOWSKI, K. (2007) Experimental studies on specific cutting resistance while cutting with narrow-kerf saws. Advances in Manufacturing Science and Technology. 31(1): 49–63. 4. AGAPOV, A. I. (1983) Dinamika processa pilenija drevesiny na lesopil´nych ramach. (In Russian: Dynamics of wood sawing on frame sawing machines). Kirovskij Politechničeskij Institut, Izdanije GGU, Gor´kij. 5. ORLICZ, T. (1988) Obróbka drewna narzędziami tnącymi. (In Polish) Skrypty SGGW-AR w Warszawie, Wydawnictwo SGGW-AR, Warszawa. 6. ATKINS, A. G. (2009) The science and engineering of cutting. The mechanics and process of separating, scratching and puncturing biomaterials, metals and non-metals, ButterworthHeinemann is an imprint of Elsevier, Oxford. 7. GROTTE, K. H., ANTONSSON, E. K. (Eds.) (2008) Chapter 7.3: Machining Processes. pp. 606–656 In: Springer Handbook of Mechanical Engineering, Part B: Applications in Mechanical Engineering. Springer. 8. ORLOWSKI, K. A. (2010) The fundamentals of narrow-kerf sawing: the mechanics and quality of cutting. Technical University in Zvolen, Publishing house of the Technical University in Zvolen. 9. PANTEA, R.C. (1999) Wood cutting system: modelling and process simulation. Mémoire présen té à la Faculté des études supérieures de l'université Laval pour l'obtention du grade de maître ès science (M.Sc.). Département de génie mécanique FACULTÉ DES SCIENCES ET DE GENIE, UNIVERSITÉ LAVAL, (National Library of Canada). 10. ATKINS, A.G. (2003) Modelling metal cutting using modern ductile fracture mechanics: quantitative explanations for some longstanding problems. International Journal of Mechanical Sciences, 45: 373–396. 11. ORLOWSKI, K. A., OCHRYMIUK, T. (2010) The prediction method of the shear angle in the cutting zone during wood sawing. Ann. WULS-SGGW, Forestry and Wood Technology, 72: 99–102. 12. ORLOWSKI, K. A., ATKINS, A. (2007) Determination of the cutting power of the sawing process using both preliminary sawing data and modern fracture mechanics. pp. 171–174. In: Navi, P., Guidoum, A. (Eds.) Proc. of the Third Inter. Symposium on Wood Machining. Fracture Mechanics and Micromechanics of Wood and Wood Composites with regard to Wood Machining., 21–23 May, Lausanne, Switzerland. Presses Polytechniques et Universitaires Romandes, Lausanne. 13. ORLOWSKI, K.A., PAŁUBICKI, B. (2009) Recent progress in research on the cutting process of wood. A review COST Action E35 2004–2008: Wood machining – micromechanics and fracture. Holzforschung, 63: 181–185. 14. BEER, P., (2002) Obróbka skrawaniem obwodowym drewna nowo opracowanymi narzędziami (In Polish: Wood peeling with new elaborated tools). Roczniki Akademii Rolniczej w Poznaniu, Rozprawy Naukowe, Zeszyt 330. Wydawnictwo Akademii Rolniczej im. Augusta Cieszkowskiego w Poznaniu, Poznań. 15. ORLOWSKI, K. A., OCHRYMIUK, T., ATKINS, A. (2010) Specific cutting resistance while sawing of wood – the size effect. Ann. WULS-SGGW, Forestry and Wood Technology, 72: 103–107.

288


The Relationship between Macroscopic Chip Type and Microscopic Crack Behavior in Wood Cutting Koshizuka, Miho 1, Ohtani, Tadashi 2 and Inoue, Masafumi 3 1

Graduate School of Agricultural and Life Sciences, The University of Tokyo, Tokyo, Japan 2 Department of Natural Science, Tokyo Gakugei University, Tokyo, Japan 3 Asian Natural Environmental Science Center, The University of Tokyo, Tokyo, Japan

ABSTRACT The relationship between microscopic crack behavior and macroscopic chip formation when cutting parallel to the grain was examined. Cutting tests were conducted in sample specimens of Japanese cedar (Cryptomeria japonica D.Don) using an SK tool at different tip angles. The results showed that split-type chips were formed at small and medium tip angles and compressive-type chips were formed at the large tip angle. Propagation of microscopic cracks was also observed in the intercellular layer and the radial wall of the lumen. In examining the relationship between the macroscopic chip type and the microscopic crack behavior, it was clarified that the microscopic crack propagation in the compressive-type chip was controlled to be much shorter due to the fracture of mode II. It was also concluded that the so-called “preceding crack” in the split-type chip was related to the propagation of microscopic cracks in the intercellular layer.

INTRODUCTION A Chipped or torn grain is common defects on the surface texture of machined wood. These defects might be caused by the fracturing of the crack that precedes the tool (the so-called “preceding crack”). The fracture phenomena in machined wood can be observed at various levels: the macroscopic scale on the annual ring structure, the microscopic cellular scale, and the cellulosic molecular scale. Previous studies on the wood-cutting process have focused on the observation of macroscopic fracture phenomenon in cutting [1-8]. Other reports have investigated chip types in the orthogonal cutting of wood [1, 2], the relationship between the chip type and the cutting force [3], the chip formation affected by cutting speed [4] and mechanical analyses of the chip formation [5-7]. Especially, researches on split-type chips when cutting parallel to the grain have reported on the mechanism of preceding crack generation in the chip formation [8]. Moreover, research on the fracture phenomenon has also reported that fractures perpendicular to the grain could be observed on a cellular scale on the cut surface [9, 10]. In focusing on the fracture of cell walls with a cellulose molecular chain, the fractography research could analyze the fracture of a single tracheid cell wall [11]. On the other hand, the optimum design for tool machining has been developed based on a recent computer simulation technique, and one of the researches regarding the wood surface finish analyzed the stress distribution using the finite element method [12]. Also, the planning skill of

289


Japanese shrine carpenters in obtaining superior surface finish with high-quality gloss and watershedding is well known, but the empirical basis for this superiority has not been clarified scientifically. Therefore, studying the detailed fracture phenomena should advance the optimum design of surface finish and help us understand the traditional technique. However, such systemic research on the macroscopic to microscopic scale of wood cutting has not yet been conducted. In this study, both the microscopic crack behavior and macroscopic chip type were observed for cutting at different tip angles, and the relationship between the two was investigated.

MATERIALS AND METHODS Materials The sample species used was Japanese cedar (Cryptomeria japonica D.Don), the average density was 0.35 g/cm3 and the moisture content was 11.0 %. The sample test specimen was cut from a block of early wood with the dimensions of 4 (L) x 4 (T) x 5 (R) mm. Cutting test We designed the experimental setup especially for image analysis of wood cutting behavior. In the experimental setup, the test specimen was mounted on a moveable stage that allowed for minute translation. The stage also could be translated back and forth toward the cutting tool, which was fixed in the middle portion of the visual camera range. Table 1 shows the conditions for the cutting tests in this experiment. The cutting angle was set at 30, 50 and 70 degrees, and the clearance angle was 10 degrees. In the experiment, the cutting surface was selected in the tangential section of the wood, and the test was conducted in cutting parallel to the grain for a distance of 4 mm. Both the macroscopic chip formation and the microscopic crack behavior were observed from the side view of the cutting direction using the digital camera. Table 1. Condition in orthogonal cutting test. Tool material Cutting angle (°) Tool angle (°) Clearance angle (°) Cutting speed (mm/s) Depth of cut (μm)

SKH3 = 30, 50 and 70 and = 10 0.15 d = 200

Image analysis methods of cutting behavior Macroscopic chip formation was observed on digital images taken at intervals of 0.5 mm in the cutting distance from 0.5 mm to 2.0 mm. The microscopic crack behavior was also observed for intervals of 0.1 mm in the cutting distance from 0.1 mm to 0.8 mm.

290


In the observation of the microscopic crack behavior, the initial cut position in contact between the test specimen and tool was set as shown in Figure 1(a) and (b). The position was in the intercellular layer between the cells as shown in Fig. 1(a) and the radial wall of lumen in Fig. 1(b), respectively. The cutting test was conducted ten times for each cutting angle. From the experimental images, the crack propagation length was measured using the image analysis software. The propagation length (r) was measured in the first linear propagation from the tool tip, and the next crack length was measured in the case of a new propagation in a different direction. The crack propagation length (r) was calculated as the average value of ten tests.

(a)

Tool

Intercellular layer

(b) Lumen

100μm

Fig. 1. Experimental condition of cutting tool in tangential surface orthogonal cutting parallel to grain. Note: (a) Cut in intercellular layer, (b) Cut in radial wall of lumen.

RESULTS AND DISCUSSION Macroscopic chip type in parallel cutting to grain Figure 2 shows the result for macroscopic chips observed when cutting parallel to the grain with a 200-µm depth of cut (d = 200) using tools with different cutting angles. Fig. 2 shows an example of the result at the cutting distance of 0.5 mm with tool tip angles of 30, 50 and 70 degrees. In the result shown in Fig. 2, the chip is produced in front of the tool tip; large cracks are propagated for both cutting angles of 30 and 50 degrees. In other words, a “preceding crack”, as is well known in cases of cutting parallel to the grain, can be observed. Also at the propagative end of the preceding crack, the chip is produced with bending action, and several bent chips are observed on the cutting angle at 50 degrees. This chip type can be classified as the “split-type chip” reported by Franz and Walker, and was especially remarkable at the cutting angle of 50 degrees. In contrast to these results, no large crack was observed at the cutting angle of 70 degrees, and shorter chips tended to be produced. The short chip seemed to be compressed on the rake angle surface by the action of the progressing tool. This result indicates that the “preceding crack” is not produced at a large cutting angle, and that the chip type at a large cutting angle is different from that at 30 and 50 degrees. The chip type can be classified as a “compressive type chip” by Franz and Walker. From the above results, we clarified that split- and compressive-type chips were observed when cutting parallel to the grain depending on the cutting angle. Below we discuss the microscopic crack behavior observed for the above typical chip types.

291


Microscopic crack in cutting parallel to grain Figure 3 shows the results for microscopic crack behavior in the macroscopic chip formation observed when cutting parallel to the grain at different tip angles. In the case of cutting in the intercellular layer as shown by the upper images of Fig. 3, crack propagation was observed in the progress of the intercellular layer. Especially, the crack length tended to be larger with a smaller cutting angle, and the crack at the cutting angle of 70 degrees was much shorter than that at 30 and 50 degrees. Also, as shown by the lower images, the crack propagation with the cutting angles of 30 and 50 degree was observed to be on a slant to the radial wall of the lumen. Especially at a 30-degree angle, the longer crack is propagated to the radial wall of lumen. Also note that no crack in the radial wall of the lumen is observed at the 70-degree angle cut. The crack propagation length was measured in order to evaluate the microscopic crack behavior quantitatively. Table 2 shows the result of crack length in the intercellular layer and radial wall of the lumen. Each result of the upper line as shown in Table 2 indicates the cut in the intercellular layer, and the lower indicates the cut in the radial wall of lumen. The average and standard deviation of the crack lengths of ten trials were also calculated for each 0.1-mm interval along the cut. For the cutting angles of 30 and 50 degrees, the crack propagation in the intercellular layer was observed for cuts in the intercellular layer and in the radial wall of the lumen, and the crack lengths ranged from approximately 150 to 200 µm. By contrast, the crack length for the 70degree cut was about half that of the 50-degree cut; the greater angle appeared to control the crack propagation. At a cutting angle of 30 degrees, a relatively long crack in the radial wall of lumen is observed on both conditions of cut in the intercellular layer and lumen. However, cracks by cutting in the intercellular layer were not observed for either 50- or 70-degree cutting, while short cracks by cutting in the radial wall of the lumen were observed in the radial wall of the lumen. Thus, it was clarified that relatively long cracks in the intercellular layer and radial wall of lumen occur when cutting with a small tip angle, a long crack only occurs in the intercellular layer with a medium tip angle, and a much shorter cracks in the intercellular layer occur with a large tip angle. = 30°

= 50°

= 70°

1mm

= 30°

1mm

1mm

= 50°

= 70°

100µm

100µm

100µm

Fig. 2. Macroscopic chip type in different cutting angle. Note: The picture below is a low magnification, and the picture above is a high magnification.

292


 (°)

30

50

70

Tool Crack in intercellular layer 100μm

Crack in radial wall of lumen

Fig. 3. Microscopic fracture appearance in different cutting angle.

Table 2. Microscopic crack propagation length in different cutting angle.

Crack in radial wall of lumen

Crack in intercellular layer

(°)

30

50

70

Cut in intercellular layer

148.8±100.9

142.1±96.0

72.8±37.6

Cut in radial wall of lumen

148.0±95.8

183.8±97.6

78.2±43.7

Cut in intercellular layer

125.3±86.8

―

―

Cut in radial wall of lumen

133.2±82.2

70.6±47.0

22.3±6.9 Unit: µm

Note: Values are means ± standard deviations.

Relationship between microscopic crack and macroscopic chip type The following paragraph discusses the relation between the “preceding crack” phenomena observed for different macroscopic chip types and the microscopic crack behavior. Therefore, a further experiment cutting parallel to the grain with the tip angle of 50 degrees was performed at different depths of cut. Figure 4 shows the measured length of the crack propagation in the intercellular layer when cut parallel to the grain at depths from 50 to 400 µm. The black and white circles indicate when the tool tip progressed into the intercellular layer and the radial wall of lumen, respectively. At a 100-

293


µm depth of cut, cracks propagated in the intercellular layer were smaller than r = 100 µm, and no cracks were observed for cuts in the radial wall of the lumen. However, at cutting depths d ≥ 200 µm, cracks were observed in both the intercellular layer and lumen, and the crack lengths tended to increase with increasing d. The results indicate that when cutting parallel to the grain with the medium tip angle at a depth of d = 200 µm, in which the split type chip is observed, microscopic cracks always appears in the intercellular layer at the 200-µm boundary. Also, in the split-type chip, the macroscopic fracture phenomenon of a “preceding crack” was observed in cutting with the medium tip angle at the same depth.

Crack propagation length　r　（µm）

From above results, when thinking about the relation between the macroscopic chip type and microscopic crack behavior, cracks in the intercellular layer and radial wall of the lumen can be observed in the split-type chip. Apparently, the preceding crack in the split-type chip is related to microscopic cracks that develop in the intercellular layer. The above results clarify that the microscopic cracks in the intercellular layer for the compressive-type chip are much shorter due to the fracture mechanics of mode II.

1400 1200 1000 800

Crack in intercellular layer ●： Cut in intercellular layer ○： Cut in radial wall of lumen Preceding crack (in macroscopic observation)

600 400 200 0 0

100

200

300

400

500

Depth of cut　d　（µm） Fig. 4. Relation between crack propagation length and depth of cut. Note: Symbols and error bars mean the average value and standard deviation of the crack propagation length in the measurement point of each cutting distance.

CONCLUSIONS The microscopic crack behavior and macroscopic chip type formation were observed when cutting parallel to the grain by using the tool at different tip angles. In examining the relationship between the crack behavior and the chip type, the following results were obtained. (1) The macroscopic chip type when cutting parallel to the grain differed at different tip angles: split-type chips appeared at small and medium tip angles, and compressive-type chips at the large tip angle. (2) Other differences were noted in the microscopic crack behavior: relatively long cracks in the intercellular layer and radial wall of lumen appeared when cutting with a small tip angle in the

294


intercellular layer and radial wall of lumen, a long crack only in the intercellular layer appeared with a medium tip angle than that in the radial wall of lumen, and a much shorter crack appeared in the intercellular layer with a large tip angle. (3) When examining the relationship between the macroscopic chip type and the microscopic crack behavior, it was clarified that the crack propagation in the above compressive-type chip was controlled much shorter due to the fracture mechanics of mode II. (4) It appeared that the “preceding crack” in the split-type chip was related to the propagation of microscopic cracks in the intercellular layer.

REFERENCES 1.

Franz, N.C. (1955) An analysis of chip formation in wood machining, Forest Products Journal, 5 (10): 332-336.

2.

Walker, K.J.S., Goodchild, R. (1960) Theory of cutting, Forest Prod. Res. Special Report, 14: 1-25.

3.

Stewart, H.A. (1971) Chip Formation When Orthogonally Cutting Wood Against the Grain, Wood Science, 3 (4): 193-203.

4.

Inoue, H., Mori, M. (1979) Effects of Cutting Speed on Chip Formation and Cutting Resistance in Cutting of Wood Parallel to the Grain, Mokuzai Gakkaishi, 25 (1): 22-29.

5.

Triboulot, P., Asano, I., Ohta, M. (1983) An Application of Fracture Mechanics to the Wood-Cutting Process, Mokuzai Gakkaishi, 29 (2): 111-117.

6.

Tochigi, T., Tadokoro, C. (1985) Change of Cutting Stress in the Progression of the Dulling of the Tool Edge, Mokuzai Gakkaishi, 31 (11): 880-887.

7.

Atkins, A.G. (2004) Rosenhain and Sturney revisited: the ’tear’ chip in cutting interpreted in terms of modern ductile fracture mechanics, Proceedings of the Institution of Mechanical Engineers, 218 (10): 1181-1194.

8.

Kato, K., Asano, I. (1982) An Analysis of Type I Chip Formation in the Wood Cutting Process, Annual report of the Faculty of Education, Gunma University. Art, technology, health and physical education, and science of human living series, 18: 9-22

9.

Hayashi, D., Tochigi, T., Inoue, H., Ogasawara, K. (1975) Observation on Rapture Formation of Cell Walls at the Cellular Level in the Wood Cutting (0-90) Process (III), Wood Industry, 30 (12): 544-548.

10.

Ohbayashi, H., Momoi, T., Tochigi, T., Kobayashi, J. (2007) Evaluation of Machinability of Thermal-Compressed Japanese Cedar at the Cellular Level, Mokuzai Gakkaishi, 53 (5): 262-268.

11.

Furukawa, I., Saiki, H., Harada, H. (1975) Tensile Fracture Process of Single Coniferous Tracheid, Journal of the Society of Materials Science, Japan, 24 (264): 855-861.

12.

Kinoshita, N. (1983) Analysis of the Veneer-Formation Process II, Mokuzai Gakkaishi, 29 (12): 877-883.

295


Orthogonal Cutting as a Method for the Determination of Fracture Properties of Oriented Wood Tissue Merhar Miran, Bučar Dominika G., Gospodarič Bojan, Bučar Bojan University of Ljubljana, Biotechnical Faculty, Department of Wood Science and Technology, Slovenia (e-mail: [email protected])

ABSTRACT The method of determining critical stress intensity factor on the basis of the chip segment length of type I chip, originating from the process of rectilinear orthogonal cutting of beech wood (Fagus sylvatica L.) in the direction of 90°-0° is shown. Experimental cuts were made on 10 mm thick samples at the rake angle of 31° and 42°, while the chip thickness varied within the range from 0.1 to 0.3 mm. The chip formation process was recorded by a high-speed video camera with recording speed of 2000 fps. The images demonstrate chip formation as a result of fracture and chip break as a result of bending. Assuming that the crack propagates during tissue separation only when stress intensity at the crack tip is critical, and the compressive stress increases in the chip formation process up to the limit strength, we can calculate the critical stress intensity factor, i.e., determine the fracture toughness of the material on the basis of a known of representative chip segment length and compressive strength in longitudinal direction. The crack continues to propagate until compressive stress reaches the strength which occurs at the point of chip break. Using the finite element method we modelled a representative chip segment of varying length and thickness. We varied the load on the chip segment and calculated, for each case, the stress intensity at the crack tip and compressive stress in the hypothetic chip. Critical stress intensity factors determined on the basis of the representative 0.3 mm thick chip segment length correspond very well to the values obtained by conventional fracture test. There are minor deviations in the case of thinner chip segments, which are most probably the result of an increased asymmetry of the sample. The method is very simple and fast since fracture properties of the material in question can practically be determined by a single linear cut. It should be emphasized that the specimen preparation is undemanding because attention needs to be paid only to the tissue orientation.

INTRODUCTION Numerous authors dealt with the occurrence and characteristics of fractures in the past [1, 2, 3, 4, 5, 6]. Using conventional tests they determined fracture toughness of symmetrically loaded samples of prescribed shapes and dimensions with artificially caused initial crack. In the case of cutting there is a distinctive asymmetry, with a relatively thin chip on one side and, compared to the chip thickness, an almost semi-infinite space on the other side, which undoubtedly affects the value of fracture toughness as well. Examining the available literature on chip formation and/or modelling of the cutting process, we have found no research which would provide the fundamental knowledge necessary to determine the fracture properties of the studied material merely on the basis of the length of specifically shaped chip segment. When calculating the

296


modelled force, some authors did take into account the energy necessary for the formation of new surfaces [7, 8], but only for a continuous chip. Triboulot [9] considered the cleavage and shear fracture mode in chip formation modeling, but only as a criterion for the formation of a certain type of chip. In the case of oriented wood tissue the process of mechanical separation of tissue produces chips, which can after all be classified into types to a considerable extent despite a relatively great variability of relevant material properties. Classification into types can be made on the basis of morphological properties of the formed chips. During experimental cutting with variable relevant parameters, it was concluded that all chips formed by cutting longitudinally can be classified into only three significantly different shapes, marked as chip type I, II or III [10]. The type I chip is formed as a result of alternating modes of wood tissue destruction, i.e., a cleavage fracture and chip break due to bending. The contact of the cutting tool with the machined material produces deformations which result in the accumulation of elastic energy in the material. When it is sufficiently high, a morphologically suitable discontinuity in the machined material – always abounding in wood because of its specific fibrous structure – can cause an instantaneous release of this elastic energy which is reflected as a sudden, usually oriented fracture in the material. The extent of the fracture actually depends on the quantity of accumulated energy alone. Further penetration of the cutting tool causes the crack to propagate and the material to cleave until the moment when the stress in the chip under bending load reaches the critical limit, i.e. limit strength (Figure 1). As the crack propagates the compressive stress in the chip under bending continuously increases.

Figure 1: Bending and crack tip stress distribution As the type I chip formation undoubtedly involves the process of oriented tissue separation which is fatally associated with a significant instantaneous fracture, an opposite presumption can be made: namely, that given the known relevant factors it is undoubtedly possible to determine the usually complex fracture properties of the studied material by using a relatively simple cutting experiment. In fact, it is merely necessary to set up the conditions for type I chip formation and to determine the significant segment length of the formed chip.

MATERIAL AND METHODS Cutting experiment were carried out on beech samples (Fagus sylvatica L.) of dimensions 130 mm long and 10 mm thick in the 90°-0° direction. The specimen was longitudinally oriented, without visible defects, with tangential texture and density of 628 kg/m3, and equilibrium moisture content of 9.5 ± 0.5 %.

297


The specimen was fixed onto a Kistler four-component dynamometer type 9272 (Figure 2). The clamping system natural frequency was about 650 Hz in the direction of cutting, and 620 Hz perpendicularly to the direction of cutting. Forces were measured in the direction of cutting (Fx) and perpendicularly to the direction of cutting (Fy). The sampling rate was 20 kHz. Data were captured by an AT-MIO 16-E1 measurement card and LabView software by National Instruments.

Figure 2: Experimental system The rake angles of cutting were 16°, 22°, 31°, 42° in 54°. The chip thickness ranged from 0.1 mm to 0.3 mm in intervals of 0.05 mm. At the rake angles of 16° and 22° the shear type II continuous chips were formed, as mentioned by Koch [10], while type I chips were formed at other angles. The average length of type I chip segments was between 0.3 mm at chip thickness of 0.1 mm and 1.62 mm at chip thickness of 0.3 mm. The HSS cutting tool with 30° sharpness angle produced by Leitz was mounted on a guided carriage which was fed by means of a hydraulic cylinder. The cutting speed was set by a one-way flow valve allowing continuously adjustable hydraulic medium flow control. Because of the tendency to have the chip breaking frequency lower than the dynamometer’s natural frequency, we set the cutting tool speed to 30 mm/s. The cutting process was also recorded by the Olympus I-SPEED 3 high-speed video camera. The recording speed was 2000 fps with the resolution of 1280 x 1024 dpi. We used the SIGMA 105 mm F 2.8 macro lens and macro extension rings with the total length of 136 mm and an intermediate lens. This configuration reduced the frame size to 5 mm. The camera was mounted on the carriage with the cutting-tool. The chip segment lengths (L) of the formed type I chips were measured under the microscope with 20x magnification. We used a Euromex microscope type SKO 37058. The objective lens magnification was 2x, and a digital camera by Digital Eyepiece type TCA-3.0 with 10x magnification was inserted in the place of eyepiece.

298


Using a four-point bending test we also determined the modulus of elasticity El and bending strength σu for the specimens of 100 mm x 10 mm x 5 mm in size. The bending strength in longitudinal direction was also determined by a clamped cantilever specimen where the cantilever was 0.5 mm thick, 10 mm wide and 16 mm long. The critical stress intensity factor KIC for fracture mode I was determined for the cut specimen. The specimen was of CT shape, 120 mm long, 100 mm high and 10 mm thick. The crack was 55 mm long. The sample direction was TL, which means that it was loaded in tangential direction, while the crack propagated in longitudinal direction. The loading force was measured by a dynamometer on a universal testing machine, while the crack mouth opening was measured by a measuring instrument that was made for this purpose from steel gages onto which strain gages were mounted. The specimen and chip were modelled by the finite element method using the Ansys software, with plane-strain conditions and orthotropic properties of the material. The measured value of 14500 MPa was taken for the modulus of elasticity in longitudinal direction El, while the relevant data for other directions, as well as the values for shear moduli G and Poisson’s ratios were taken from literature [11]: El =14500MPa, Et = 1140 MPa, Er = 2240 MPa, νlt =0.518, νtr =0.36, νlr =0.45, Glt = 1055 MPa, Gtr = 460 MPa, and Glr = 1600 MPa. The critical stress intensity factor KIC was calculated by means of Rice's J-integral [1] ∂u (1) J = ∫ (Wnx − Ti i )ds ∂x Γ where Γ is the integration path, W is the strain energy density,

1 W = σ ijε ij , 2

(2)

ni is the normal to the integration path, Ti = σ ij n j is the traction vector, ui is the displacement vector, and ds is the differential of the path Γ. Assuming that J = G, where G is the strain energy release rate, and that the crack surface is parallel to the main axis of the material, the relation between GI and KI [3] can be applied

K I2 GI = ` E

(3)

where E` is the equivalent modulus of elasticity ⎡b b E = ⎢ 11 22 ⎣⎢ 2 `

⎛ b22 2b12 + b66 ⎜ ⎜ b + 2b 11 ⎝ 11

⎞⎤ ⎟⎥ ⎟ ⎠⎦⎥

−1 / 2

In case of a plane-strain deformation state, we must take into account a i3 a j3 bij = aij − (i, j = 1,2,...,6) a33

299

(4)

(5)


where constants aij are coefficients of deformability of the material a11 =

ν 1 1 1 , a2 = , a12 = a 21 = − 12 , a 66 = E11 E 22 E11 G12

(6)

The chip segment lengths ranged from 0.4 mm to 2 mm, and the chip thicknesses from 0.15 mm to 0.3 mm. For each case of thickness h and length L the end of each chip was gradually loaded with perpendicular force, and the stress intensity factor KI at the crack tip and compressive stress σx in longitudinal direction were calculated for fracture mode I by means of the J-integral. From these calculations we thus obtained the relations between the force of loading, stress intensity factor and compressive stress for chips segments of specific thickness and length. On the basis of relations obtained at a constant chip thickness we calculated, for different chip lengths, the force for the determined compressive stress in a chip, and from this force we calculated the stress intensity factor.

RESULTS AND DISCUSSION The Figure 3a clearly shows the locations of chip break and the lengths L of chip segments. The distribution of chip segment length is shown in Figure 3b. Both figures indicate considerable variability in chip segment lengths, ranging from 0.73 mm to almost 3 mm. Average segment length value is 1.62 mm, with standard deviation of 0.53 mm. The reason for such great segment length variability lies in the variability of the critical stress intensity factor, modulus of elasticity, as well as bending strength of the material, as will be shown below. Average chip segment lengths at other thicknesses were decreasing linearly with chip thickness.

Figure 3: Type I chip with segment length L, chip thickness 0.3 mm, rake angle 31° (a) and distribution of chip segment length (b)

Figure 4 shows the forces measured in the direction of cutting (Fx) and perpendicularly to the direction of cutting (Fy) together with the positions of chip breaks. The figure shows very good agreement between the position of chip break and force peak when the cutting-tool penetrates the still intact wood tissue. Minor deviations between force peak and chip break position can be seen at places – this being the consequence of lateral measurement of the chip break position. Regarding that the chip thickness was 0.3 mm, and chip width 10 mm, the break could occur

300


250

15

200

10

150

5

100

0

50

-5

0 2

2.1

2.2

2.3

Time (s)

2.4 Fx

-F y (N)

Fx (N)

sooner on the side of the chip where break position was measured by a microscope than on the remaining width of the chip, or even later. Thus it was possible to measure a shorter or longer chip segment length than its average length.

-10 2.5 -Fy

Figure 4: Cutting force components Fx and negative Fy together with chip segment length (dashed lines); chip thickness 0.3 mm; rake angle 31°

Forces in the direction of cutting and perpendicular to the direction of cutting vary significantly with respect to time as a consequence of the alternating effects of the fracture and chip break due to bending. Figure 5 shows a detailed history of cutting force components, and Figure 6 shows the chronology of a chip segment formation. When the cutting-tool comes in contact with the still intact wood tissue, the cutting-tool tip is first impressed in the wood tissue in which elastic energy accumulates (Figure 5-point a and Figure 6a). Just before the onset of crack the cutting force components Fx and Fy are the highest (Figure 5-point b and Figure 6b). When the amount of accumulated energy is sufficient and the smallest crack is initiated – because of discontinuity in the material they are always present – the energy is released and a chip of type I is being formed in such a way that the crack propagates fast (Figure 5-point c and Figure 6c) to some equilibrium length (Figure 5-point d and Figure 6d). Then the cutting-tool continues to lift the chip and the crack propagates as long as the stress intensity factor at the crack tip equals critical stress intensity factor KI = KIC and as long as the compressive stress is lower than the strength σ x < σ u . As the crack grows, compressive stress increases. When it reaches the limit strength, the chip breaks (Figure 5-point e and Figure 6e) to form a segment of a certain length. At the point of break, the stress intensity at the crack tip thus equals the critical stress intensity factor K I = K IC , and the compressive stress on the upper side of the chip equals limit strength σ x = σ u . The cutting-tool continues to bend the already broken chip (Figure 5-point f and Figure 6f) all the way to the crack tip where the entire cycle is repeated. If we know the compressive strengths and the lengths of chip segments, we can determine the critical stress intensity factor KIC for fracture mode I.

301


160

b

140

c

Fx (N)

120 d

100 80 60

a

e f

40 20 2.715

2.725

2.735

2.745

2.755

2.765

Time (s)

Figure 5: Cutting force component Fx

Figure 6: Chip formation process; sampling rate 2000 fps

Average modulus of elasticity El in longitudinal direction amounts to 14500 MPa with standard deviation of 1415 MPa, compressive strengths σu obtained from four-point bending test amount to 150 MPa with standard deviation of 11 MPa, and compressive strengths of clamped cantilever specimens amount to 185 MPa with standard deviation of 12 MPa. The difference in strength between four-point bending test and clamped cantilever specimen is attributed to the fact that in the case of four-point test the specimen was thicker and represented the material average, while in the case of cantilever where the specimen was only 0.5 mm thick, we came across a more stiffer part of material. The applicable standard for bending strength determination lays down the use of four-point bending test in order to avoid the effect of shear, but in our case the cantilever clamping represents a more realistic model. The determined average value for critical stress intensity factor amounts to 0.60 MPa m with standard deviation of 0.09 MPa m .

302


Figure 7a shows the calculated relations between the chip segment length and stress intensity factor at constant compressive stress for 0.3 mm thick chip formed by cutting at a rake angle of 31°. The calculated critical stress intensity factor for the mentioned chip thus ranges between 0.5 MPa m at compressive strength 140 MPa and 0.65 MPa m at compressive strength of 180 MPa. Regarding that the bending strengths measured by four-point bending test or clamped cantilever specimen amounted to 151 MPa or 185 MPa, respectively, and in view of the measured critical stress intensity factors of the CT specimen where the values ranged from 0.52 MPa m to 0.73 MPa m with average value 0.60 MPa m and standard deviation 0.09 MPa m , the values for the critical stress intensity factor obtained from the length of chip segments are by all means satisfactory. If the variability of chip segment lengths and bending strength variability are also taken into account, the range of calculated stress intensity factor is further increased, but still remains within the range of values obtained from the static test for determining the critical stress intensity factor between 0.52 MPa m and 0.73 MPa m . Figure 7b shows the calculated critical stress intensity factor ranges for various chip thicknesses or their average segment lengths within the compressive strength range from 140 MPa to 180 MPa. The figure shows that the calculated critical stress intensity factor range decreases with the decreasing chip thickness. The reason for that is attributed to increased asymmetry in the case of thinner chips compared to thicker chips. Regarding that the calculated critical stress intensity factor at chip thickness of 0.3 mm is comparable to the values of critical stress intensity factor determined by the conventional fracture test, while the calculated critical stress intensity values in the case of thinner chips are lower, a thicker chip is by all means more suitable for determining the critical stress intensity factor since the distinctive specimen asymmetry is thus somewhat reduced.

Figure 7: Relation between chip segment length and critical stress intensity factor for bending strength between 140 MPa and 180 MPa for chip thickness 0.3 mm (a) and for diferent chip thicnesses (b). Horizontal lines represent average chip segment length.

CONCLUSION The method of determining the critical stress intensity factor on the basis of chip segment length of type I chip according to Koch [10], which originates from orthogonal linear cutting is shown. The method has proven to be suitable since the average critical stress intensity factor and its variability can be determined for the specimen by means of a single orthogonal cut. It should be

303


emphasized that specimen preparation is simple and undemanding, because we need no specimen of prescribed shapes and dimensions as is the case in conventional fracture tests. Also crack initiation which is of high importance in the conventional fracture tests, presents no problem in our method, because the critical stress intensity factor is determined solely on the basis of chip segment lengths which are not influenced by crack initiation. It is also relatively simple to conduct this experiment as opposed to the conventional fracture test where not only force but also specimen opening displacement must be measured. In the case of 0.3 mm thick chip the range of calculated critical stress intensity factor agrees very well with the values obtained from the conventional fracture test, while in the case of thinner chips the range is somewhat lower. The reason for that is attributed to a more pronounced asymmetry of thinner chips. The method is simple to use if we know certain material properties, i.e., modulus of elasticity and bending strength. If we do not know them, they should be determined experimentally, which can be deemed as a deficiency of the presented method.

REFERENCES . 1. Banks-Sills, L., Hershkovitz, I., Wawrzynek, P. A., Eliasi, R., Ingraffea, A. R. (2005) Methods for calculating stress intensity factors in anisotropic material: Part I-z = 0 is a symmetric plane. Engineering Fracture Mechanics, 72 (15):2328-2358. 2. Le-Ngoc L., McCallion H. (1997) On the fracture toughness of orthotropic materials. Engineering Fracture Mechanics, 58 (4): 355-362. 3. Sih, G. C., Paris, P. C., Irwin, G. R. (1965) On cracks in rectilinearly anisotropic bodies. International Journal of Fracture Mechanics, 1 (3): 189-203. 4. Stanzl-Tschegg, S. E., Tan, D. M., Tschegg, E. K. (1995) New splitting method for wood fracture characterization. Wood Science and Technology, 29 (1): 31-50 5. Vasic, S., Stanzl-Tschegg, S. (2007) Experimental and numerical investigation of wood fracture mechanics at different humidity levels. Holzforschung, 61 (4): 367-374. 6. Yeh, B., Schniewind A. P. (1992) Elasto-plastic fracture mechanics of wood using the Jintegral method. Wood and Fibre Science, 24 (3): 364-376. 7. Atkins, A. G. (2003) Modelling metal cutting using modern ductile fracture mechanics: quantitative explanations for some longstanding problems. International Journal of Mechanical Sciences, 45 (2): 373-396. 8. Williams, J. G. (1998) Friction and plasticity effects in wedge splitting and cutting fracture tests. Journal of Material Science, 33 (22): 5351-5357. 9. Triboulot, P., Asano, I., Ohta, M. (1983) An application of fracture mechanics to the woodcutting process. Mokuzai Gakkaishi, 29(2): 111-117. 10. Koch, P. (1985) Utilization of hardwoods growing on southern pine sites. Agriculture handbook no. 605. U.S. Department of Agriculture, Forest Service, Washington. 11. Kollmann, F. F. P., Cote, W. A. (1984) Principles of Wood Science and Technology, Volume I: Solid Wood. Berlin, Springer-Verlag.

304


Machining of Wood using a Rip Tooth: Effects of Work-piece Variations on Cutting Mechanics Naylor, Andrew.1*

Hackney, Philip.1

Clahr, Emil.2

1

School of Computing, Engineering and Information Sciences, Ellison Building, Northumbria University, Newcastle upon Tyne, NE1 8ST, United Kingdom 2 Research and Development Centre for Wood Working, SNA Europe, Box 1103 82113 Bollnäs, Sweden

ABSTRACT Genetics and environmental conditions during the growth of wood are known to affect the intrinsic characteristics influencing cutting mechanics. To evaluate this, a full factorial experiment has been performed investigating the effects of three significant factors involved in wood machining; wood species, moisture content and grain direction. A variety of woods were evaluated (five softwood and three hardwood species) at four moisture levels. As all woods are heterogeneous, anisotropic materials, machining was performed parallel and perpendicular to the grain direction. A three axis CNC router was used to drive a tool resembling a rip tooth, at low velocity, through each of the sixty-four wooden work-piece variations at three different depths of cut. To collect quantitative data, a piezoelectric dynamometer was used with a data acquisition system to measure and record the cutting and thrust force components acting on the tool. Chip formation and work-piece deformation were observed using images taken from an optical microscope. This paper compares the published results [1-7] for planing operations with findings from the rip tooth experiment.

INTRODUCTION Research performed into optimum wood machining conditions [1, 2] suggests that there are three significant types of factor that affect the cutting mechanics: 1. Factors attributed to the machining process 2. The species of the wood 3. The moisture content of the wood Wood has three orthogonal planes of symmetry; axial, radial and tangential. Corresponding to these planes of symmetry are several different cutting directions by which different machining processes can be described. When referring to a machining direction the nomenclature states a labelling system consisting of two numbers. The first number denotes the orientation of the cutting edge to the wood grain direction; the second number denotes the movement of the tool with respect to the grain direction. To illustrate this, the three main cutting directions as shown in figure 1.

*Corresponding Author

Tel: Email:

+44 (0) 1912273624 [email protected]

305


Figure 1 – Machining directions with respect to wood grain

  

90°-90° - The axial plane or the wood end grain. Both the cutting edge and tool movement are perpendicular to the grain. 0°-90° - The radial and tangential planes, planes, cutting across the grain. The cutting edge is parallel to the grain but the tool movement is perpendicular. 90°-0° - The longitudinal plane, cutting along the grain. The cutting edge is perpendicular to the grain but the tool movement is parallel.

Evidence from fundamental literature [3, 4] suggests that cutting velocity has negligible effect on the forces acting on the tool. This is for the ranges of 0.2 mm/s – 6.3 m/s along the grain and 2.5 m/s – 50 m/s across the grain. A Review of Planing Operations Kivmaa [3] used Finish birch in a study investigating the geometric factors of the tool on cutting performance and found that the main cutting force was w inverselyy proportional to the sharpness of the tool. It is also stated at this point that the thrust force is caused by contact between the rake face and the chip. The larger the rake angle the thicker the chip and hence the lower the thrust force. This is because the chip is not being compressed. Although it is observed that there is no significant effect of cutting velocity on the major cutting force, the orientation of the tool with respect to the grain does have a significant effect on the cutting forces. For tthis planing scenario, the he highest cutting forces are observed to be in the 90°-90° 90° 90° direction with the lowest cutting forces in the 90°-0° 0° direction (cutting along the grain). In the same study, the tool sharpness and rake angle remain constant for the testing testing of 21 different species of wood. Analysis of data found a linear trend between the density of the wood and the major cutting force. From this empirical data a predictive model for cutting force was created. Extensive xtensive work into the chip formation produced through varied cutting conditions has been carried out by Franz [4, 5],, McKenzie [6], Woodson and Koch [7].. The cutting tools used in the experiments represent a wood plane that removes material across the entire width of the work workpiece. Regarding machining in the 90°-0° 90 direction (along the he grain) it was found that llarge rake angles result in negative thrust forces (acting upwards relative to the work-piece) work piece). The wood fibres split ahead of the tool and finally fail due in tension.. This type of chip is beneficial where quick removal of material rial is required. Continuous chip formation is achieved when using a very sharp tool edge and a diagonal plane of shear, shear providing an excellent lent surface finish to the work work-piece. This process can be described as ongoing shear. Dull ull tool edges, and very smal small or negative rake

306


angles cause a fuzzy chip. It is also suggested that very large depths of cut may form this chip where there is too much contact with the blade surface. This type of chip causes a raised fuzzy grain where wood fibres become protruded, hence producing a poor surface finish. An investigation into the mechanics of cutting across the grain (0°-90°) considers the veneer peeling process as a case study. This process uses high rake angles (approximately 70°) and small depths of cut (less than 1 mm). The material removal in veneer peeling is described as an ongoing shearing process initiated by a tear in compression perpendicular to the grain. McKenzie [6] investigated the effects of cutting in the 90°-90° direction. In general the cutting mechanics specify a tensile failure mode causing parallel gaps to propagate along the grain. It is noted that these gaps become larger as the moisture content decreases. Cutting forces in this direction are strongly affected by cell type, moisture content, depth of cut, and rake angle A Review of Single Tooth Operations The limited research performed on the effects of single point cutting tools focuses on the optimisation of cutting conditions for industrial sawing processes. From the available literature [811] it is apparent that the responses desired from experimentation are the forces along the major cutting edge. Chip formation is not heavily investigated. Machining in the 90°-90° direction, Axelsson [8-10] developed the prior knowledge of the machining process obtained using planing operations by investigating the effects on cutting mechanics using single point cutting tools. For sawing processes, the tool used has a side clearance of 1 mm either side to represent the set of a saw-blade. Using computerised tomography (CT) a linear relationship between the density of the wood for a specified tool path and the cutting forces was established. This linear relationship is clearly shown when cutting through a knot of much higher density to the un-defected wood-grain. Interesting results were produced from research into the effects of changing the rake angle of bandsaw teeth, machining wood in the 90°-90° direction [11]. Three teeth with 25°, 30° and 35° rake angles were examined, it was found that the largest rake angle produced the lowest cutting forces and the smallest rake angle produced the largest cutting forces. Initially, it appeared that the 25° and 35° rake angles produced a smooth work-piece finish after machining, whilst the 30° rake angle produced a rough finish with fuzzy grain. Microscope images showed that the 25° rake angle only appeared smooth when in fact the machining caused fuzzy grain which was then compressed due to the low rake angle of the tooth.

METHODOLOGY Test Equipment The experimental test rig comprised of a cutting tool driven by a 3 axis CNC router machine. The work-piece was mounted on a force dynamometer equipped with piezoelectric load cells measuring the cutting, thrust and side force components acting on the tool. Only the cutting and thrust force components were taken into consideration for this analysis. The test rig schematic diagram (figure 2) details the set-up of the data acquisition system. To obtain tool force data, the cutting tool (1)

307


was used to machine through the work-piece work attached to the force dynamometer (2). The three piezoelectric transducers in the dynamometer each generate a charge in response to the cutting forces (3.9 pC/1 N in X and Y directions, 1.95 pC/1 N in the Z direction). direction). These signals feed into the charge amplifier (3) wheree the signals are calibrated for a 10V input to the data acquisition PLC (4) (3900 3900 pC/10 V in X and Y directions, 1950 pC/10 V in Z direction: Hence 1 N = 0.01 V) V). The PLC converts the signals from analogue to digital and the data can be analysed using appropriate software.

Figure 2 - Test rig schematic with data acquisition system

Experimental design The tool used in the experiment has geometry similar to the rip tooth formation (figure 3). The tool has an orthogonal cutting edge with a width width of 1 mm and a rake angle of zero. To ensure that the cutting edge was sharp the tool was sharpened using precision grinding equipment prior to performing the test runs. The two machining directions selected for the experiment were 90°-0° (along the grain) and 0°-90° 90° (across (a the grain) as these are deemed to be the most common directions for manual wood-sawing sawing. Eight species of wood where evaluated in the experiment, five softwoods (Scots Pine, Yellow Pine, Siberian Larch, Douglas Fir and Western Wester Red Cedar) and three hardwoods (Ash, Beech and Sapele). Each of these wood species had four separate moisture levels; Dry ( +10

Deviation from target diameter interval, mm

Fig. 4

48 %

Actual top diameter distribution for (pre-)sorted logs. Target diameter interval is typically one tenth of actual diameter. Results from re-scaling 840 logs, all sizes. Red line indicates splitting the actual interval in two, the dual mini series. 65 %

Main yield

46 %

Total yield = main yield + side boards

60 % 44 % 42 %

55 %

40 % 50 % 38 % 36 %

45 %

After

After

34 %

Prior

Prior

40 %

32 % 30 %

35 % 10

20

Fig. 5

30 Diameter, cm

40

10

20

30 Diameter, cm

40

Main yield and total yield both increased after investment.

This strategy has proved successful. As can be readily seen from Fig. 5, main as well as total yield has increased. For the centre yield, the most valuable boards, the increase is most pronounced for medium-sized logs. These are the logs with most varying options for breakdown. It should not be forgotten, however, that there is a downside: Both planning and follow-up of logistics is far more complicated when sawing dual orders. Applying the analytic-empirical regression approach [1, 2], sawn timber recovery improved by 2 348 m3 (centre boards 1 020 m3, side boards 1 328 m3), equivalent to 2.0% of annual roundwood consumption and corresponding to 2.3% increase in the sawmill's added value. This is the

358


main contribution to increasing the sawmill's contribution value and making the investment a profitable one. Timber Quality Prediction in 3D Scanner Automated timber quality prediction has proven valuable applying 2D shadow scanners on unbarked log, and in particular for medium-sized logs from ca. 20 to ca. 30 cm top diameter. The prediction has been based on sweep (to avoid compression wood) and taper (to avoid large knots). It was hypothesised that quality prediction would improve when applying a 3D scanner on barked logs. However, no good model for this new intake has so far been identified. One reason might be quality differences between twin boards from the same log, as illustrated in Fig. 6. New approaches are now being sought. 60 Green knots, max diameter in mm

Left centre board

50 40 30 20 10 0 0

10

20

30

40

50

60

Right centre board

Fig 6

Green knot diameter in twin boards from the same log. Observations off the main diagonal indicate that knot size differs between the two boards. From [4].

Breakdown Capacity The breakdown capacity has undoubtedly increased. However, new bottlenecks emerged that needed to be handled – and at additional cost. The in-feed equipment for the second sawing machine group has been renewed. The most serious bottleneck now is kiln capacity. At the time of deciding the new log in-take, the assumption was to produce an unchanged volume of sawn timber, consequently reducing sawlog input and production time. After experiencing the successful operation of the new equipment, however, this was changed to producing at full capacity the same weekly hours as before, considerably increasing the weekly sawn timber volume – and running into a kiln bottleneck. At the time of writing, the solution for this has not been decided. Also, labour cost and maintenance have been reduced. However, some additional cost to finetune all new equipment had to be accepted.

359


CONCLUSION Summing up the main results:       

The 3D scanner applied after debarking substantially improved scaling accuracy Care should be taken to prevent the debarker from damaging the logs The dual order mini-series strategy has proven successful for this case sawmill Centre board and side board recovery both increased significantly Planning and follow-up is more complex with dual simultaneous orders No good model for timber quality prediction has been identified Investment in the chosen log intake plant has proven profitable

Acknowledgement This report is based on analyses made at the sawmill Moelven Numedal As (www.moelven.com) in close cooperation with production manager Bjarne Hamar. The work is partially funded by Norwegian authorities through SkatteFunn.

REFERENCES 1.

Gjerdrum, P. (2007a) A combined analytic-empiric approach for modelling sawn timber yield. In Acker, V. J; Usenius, A. (eds.) Modelling the wood chain: Forestry – wood industry – wood product markets. Proceedings from COST E44, Helsinki, September 17-19, 2007. ISBN 9789080656536. pp. 49-57

2.

Gjerdrum, P. & Hamar, B. (2010) High capacity timber sawing based on accurate log scaling. In: Meier, P. (ed.): Nordic-Baltic Network in Wood Material Science and Engineering (WSE). Proceedings of the 6th meeting, October 21-22, 2010, Tallinn, Estonia. ISBN 978-9949-23-033-4. pp. 101-107

3.

Gjerdrum, P. (1998) Målenøyaktighet og repeterbarhet. Forestia FoU, 9 p. (in Norwegian, restricted)

4.

Gjerdrum, P. (2007b) Analyse av skurutbytte og kvalitetsutfall ved Moelven Numedal. Oppdragsrapport Skog og landskap 06/2007, 32 p. (in Norwegian)

360

Component Production Poster Presentations


Bondability Study of Three Guianese Woods for Glulam Manufacturing Bourreau D.1, Beauchêne J.1, Aimene Y.1, Nait-Rabah O. 1, Thibaut B.2 1

2

UMR Ecofog, Université Antilles Guyane, Cayenne FRENCH. GUIANA. CNRS Laboratoire de Mécanique et Génie Civil, UMR-5508, Montpellier -FRANCE.

ABSTRACT To promote local wood in construction, bondability of tropical hardwood was performed to attempt gluing conditions parameters for an industrial manufacturing in local climate. Unfortunately, due to their specificities (dense wood species, high presence of removals, high shrinkage coefficients…), tropical hardwoods are known to be difficult wood species to be glued, and tropical conditions added some difficulties (high temperature and high moisture). So, those species are not adequate for this purpose. This study underlines different gluing parameters which influence delamination tests. Results show that dense wood species can be glued for structural purposes, sometimes, easier than wood with medium specific gravity. They also shown the aptitude to perform glulam for some Guianese species, and give some recommendations and gluing parameters.

INTRODUCTION Invented in 1906 per Otto Hetzer, glued-laminated timber is nowadays well-known in temperate countries as a high performed composite engineering material used in construction. It is made with several wood lamellas fitted together by glue. It is an ecological and economical product that presents high mechanical properties, and behaves well to fire, chemical and biological attacks comparing to solid wood [1]. In 2010, glulam importation has extensively augmented in French Guiana, due to a demographic increase. However, due to an end-use in equatorial conditions and despite the existence of a large Amazonian forest, glulam beams imported need to attempt specifies requirements due to severe fungi and termite decades. Gluing successfully wood depends on how well we understand and control the complexity of factors that constitute the individual links in the adhesive bonding of wood components. Moreover, in equatorial country, the principal problematic of its conception is, in one way the high temperature and humidity which can disqualify the adhesive [2] and in the other way the tropical hardwood properties such as high specific gravity which is highly correlated to wood porosity, high shrinkage coefficient, presence of removals… [3]. All these factors require an adequate selection of wood species. The gluing step of conception is the most critical point on the manufacturing process of glulam. Indeed, creation of a structural bonded timber requires adequate settlement between gluing step parameters under tropical conditions, appropriate wood and moisture content, careful surface preparation of the lamellas before gluing, homogeneous glue joint thickness, sufficient penetration of adhesive in lamellas …. An adequate choice of all these parameters can ensure wood specimens to pass successfully validation tests [4] 1 .

1

Notice that wood wettability was not performed because of missing test material.

361


However for these requirements, performed studies [5, 6, 7] on the gluing of tropical hardwoods, show that, first, sanding do not increases the gluing resistance, due to cell damage (compressed cells, crushed or partly cut fibres)[8] and to surface contamination by dust, plugging vessels or lumens [5]. Secondly, authors show that too high pressure is not benefit to glue high density wood. Working on Azobé (Lophira alata), they reached an optimum of pressure around 0.7 MPa, upper pressure decreases mechanical resistance of glue joint [6]. Last but not least, they insist on the planning step which is recommended to be done after machining, in the same working day [7]. Indeed, after planning, timber surface will attract gaseous molecules and fine particles increasing the amount of surface contamination reducing bond fracture toughness [8, 9]. Ideally, the resin should penetrate wood surface, filling voids caused by pores and vessels, checks and other anatomical features, emerging from surface preparation. At the end of the curing time, and if the adhesive was fluid enough to penetrate correctly in adherents, resin is locked in voids and high energy is needed to separate lamellas [9].

MATERIALS AND METHODS Due to a large diversity of wood species in French Guiana, industrial constraints and physical wood properties imposed a rigorous choose of species. Three wood species were selected: Qualea rosea [10], Dicorynia guianensis [11] and Peltogyne venosa [12].Principal wood properties are given in table 1. Table 1: Wood properties [10, 11, 12] Latin Name

Sg

Qualea rosea Aubl. Dicorynia guianensis Ashm. Peltogyne venosa (Vahl) Benth.

0.73 0.76 0.85

Rb (%) 17.3 16.3 13

Rt (%) 10 9 7.5

Rr (%) 6 5.4 5

V/mm² 3-8 1-2 3-7

Vdiam (µm) 140-190 225-300 120-210

I

UC.

2 3 3

2 3 3

With: Sg = Specific Gravity; Rb=Total shrinkage coeff.; Rt=Tangential shrinkage coeff.; Rr=Radial shrinkage coeff.; V/m²= number of vessel per square meter; Vdiam= Mean vessels diameter; I= impregnation and UC=Use Classes.

Thus, these parameters defines those woods as dense species (specific gravity upper than 0,7). Concerning Q. rosea and D. guianensis, their shrinkage coefficients are high and induced severe variations during drying. Moreover, due to a good impregnation and sufficient large vessels, Q. rosea is more porous than the other species, whereas P. venosa shows carbonization during manufacturing. Those observations define these wood species as “difficult to be glued”, and function of their own properties, recommendations (table 2) could enhance bondability of those hardwoods [13].

362


Table 2: Requirements to increase bondability of wood function of some wood properties [13] Requirements

Dense1

Porous2

High shrinkage3

Carbonization4

Q, D, P

Q

Q, D

P

Spread glue on both faces needed  to be assembly  Use a higher glue viscosity  Have two gluing step   Increase glue amount Increase temperature during the  gluing step   Apply a higher pressure   Rabbet longitudinal face   Sand surfaces prior to glue Decrease the transversal section   of lamellas Notice that: Q= Qualea rosea, D= Dicorynia guianensis and P= Peltogyne venosa

The resin used for the gluing step is a Resorcinol-Phenol-Formaldehyde type (Prefere 4094), available on the market according to European standards. It was obtained by mixing 20 parts of hardener and 100 parts of resin. In order to validate a structural wood gluing assembly, designed for construction end-uses, two tests are performed on standard specimens in order to guarantee good mechanical resistance of glulam. The most pertinent one is the delamination test [14] which consists in the appreciation of the glue joint resistance after ageing cycles simulated by 2 successive cycles consisting of water immersion under pressure and drying of specimen. This test induces severe hygroscopic variations. It is conclusive if the delaminate factor D (D is the lengths’ ratio of measured delamination and the total glue joints) is lower than 10%. D is given by:

(1) To performed glulam specimens in accordance with this standard, some boards were planned and sawn in dimensions 28x100x700 mm³. Three lamellas are then glued (the resin is spread on both faces of lamellas), asserted together during a closed assembly time (cat) and then pressed with a manual screw press for at least, six hours. Obtained beams are air stabilized for one week. Then, they are sawn-cut in 90×95×75mm³ samples, homogenously distributed, intended to delamination test. Two gluing parameters are considered (table 3). Each parameter is tested on two beam’s assemblies: one with three boards glued with same transverse sawn direction and the other with different fiber orientation such as quarter sawn/flat sawn/quarter sawn. Table 3: Gluing parameters tested for the bondability of three Guianese wood species Gluing parameters tested 5 Closed assembly time (cat) (min) 0.4 Pressure (MPa)

363

10 0.7

20 1


The glue amount spread is also an important parameter to take into account. Different quantities were used for specimens manufacturing, but they are presented with results because of their measurement after two weeks of stabilisation to let the glue hardening.

RESULTS The gluing step is very important in the glulam manufacturing. Different parameters are considered: the glue amount spread, the close assembly time and the assembly pressure. Theses parameters are closely dependant and are affected by environmental conditions (temperature and humidity). In that case, presenting results sorted by “theoretical” glue amount can sketch some trends for the analyze, however those can be highlighting and analyzed in another way when results are sorted by the glue amount on the assembly, measured by digital microscope. Thus, for analyzing those results, this study focuses on the rectified parameter. Influence of glue amount Sorted by the glue amount deposit on lamellas’ surfaces, delamination results for the three Guianese wood species are presented by the figure 1. First trend observed is that, more glue is put on lamellas’ surfaces, better is the joint resistance towards delamination test. Indeed, despite some conclusive results for small amount of glue, this graph highlights a limit (180g/m²) beyond which one results are conclusive. Moreover, references samples, made in Larix decidua, were tested in order to record their delamination resistance. These specimens show no delamination for a gluing step upper than 250 g/m².

Figure 1 : Influence of glue amount on delamination

This figure also shows that, in case of porous media (Q. rosea), glue amount recorded are lower than other species and delamination results are not conclusive enough to perform glue laminated timber. This observation is clearly in accordance with some recommendations presented before, advising to increase glue amount per surface unit in order to avoid thin joint (