Proceedings of National Conference on Advancements and Futuristic Trends in Mechanical Engineering; Department of Mechanical Engineering, PEC University of Technology, Chandigarh on 17th-18th Oct. 2014
ROOT CAUSE ANALYSIS OF DEFECTIVES OF A MANUFACTURING INDUSTRY
Lalit Kumar1 and D. R. Prajapati 2 PG student, Assistant Professor, Department of Mechanical Engineering, PEC University of Technology, Chandigarh, India 2 E-mail:
[email protected]
1
2
ABSTRACT Quality can also be linked to customer satisfaction. Some companies have used that definition for years, but there is now a broad move toward defining quality as total customer satisfaction. Pareto analysis is a statistical technique in decision making used for selection of a limited number of tasks that produce significant overall effect. It uses the Pareto Principle –Few causes account for most of effects. From the study, it is found that five defects contribute to more that 75% of the total defects occurring in the industry. With the help of Pareto analysis main defects affecting the company’s products’ quality are identified. Key words: Ishikawa diagram, Pareto analysis, manufacturing industry
1. Introduction Statistical process control (SPC) is a procedure in which data is collected, organized, analysed, and interpreted so that a process can be maintained at its present level of quality or improved to a higher level of quality. It can be applied wherever work is being done. Initially it was applied to just production processes, but it has evolved to the point where it is applied to any work situation where data can be gathered. Acceptance sampling is the process of randomly inspecting a sample of goods and deciding whether to accept the entire lot based on the results. Acceptance sampling determines whether a batch of goods should be accepted or rejected. The tools in each of these categories provide different types of information for use in analyzing quality. Descriptive statistics are used to describe certain quality characteristics, such as the central tendency and variability of observed data. Although descriptions of certain characteristics are helpful, they are not enough to help us evaluate whether there is a problem with quality. Acceptance sampling can help us do this. Acceptance sampling helps us decide whether desirable quality has been achieved for a batch of products, and whether to accept or reject the items produced. Although this information is helpful in making the quality acceptance decision after the product has been produced, it does not help us identify and catch a quality problem during the production process. For this we need tools in the statistical process control (SPC) category. All three of these statistical quality control categories are helpful in measuring and evaluating the quality of products or services. However, statistical process control (SPC) tools are used most frequently because they identify quality Lalit Kumar and D. R. Prajapati
problems during the production process. For this reason, we will devote most of the chapter to this category of tools. The quality control tools we will be learning about do not only measure the value of a quality characteristic. They also help us identify a change or variation in some quality characteristic of the product or process. We will first see what types of variation we can observe when measuring quality. Then we will be able to identify specific tools used for measuring this variation. Variation in the production process leads to quality defects and lack of product consistency. The Intel Corporation, the world’s largest and most profitable manufacturer of microprocessors understands this. Therefore, Intel has implemented a program it calls “copy-exactly” at all its manufacturing facilities. The idea is that regardless of whether the chips are made in Arizona, New Mexico, Ireland, or any of its other plants, they are made in exactly the same way. 1.1 The Basic Tools for Statistical Process Control (SPC) Various SPC tools are discussed in this section. 1.1.1 Flowchart The entire process is diagrammed from start to finish with each step of the process clearly indicated. All involved in the process should know their position on the flow chart and at least a partial upstream and downstream trace from their positions. All should know who their suppliers are and who their customers are in the process flow.
Page 64
Proceedings of National Conference on Advancements and Futuristic Trends in Mechanical Engineering; Department of Mechanical Engineering, PEC University of Technology, Chandigarh on 17th-18th Oct. 2014 1.1.2 Pareto Chart Pareto analysis is a statistical technique in decision making used for selection of a limited number of tasks that produce significant overall effect. It uses the Pareto Principle –Few causes account for most of effects. The idea is that by doing 20% of work, 80% of the advantage of doing the entire job can be generated. Or in terms of quality improvement, a large majority of problems (80%) are produced by a few key causes (20%). Pareto analysis is a formal technique useful where many possible courses of action are competing for attention. This technique helps to identify the top 20% of causes that need to be addressed to resolve the 80% of the problems. Once the top 20% of the causes
are identified, then tools like the Ishikawa diagram or Fish-bone Analysis can be used to identify the root causes of the problems. 1.1.3 Check Sheet A data-gathering sheet is prepared that categorized problems or defects. Check sheet information may be put on a Pareto chart or if a time analyses is included, may be used to investigate problem trend over time. 1.1.4 Cause and Effect Diagram A problem (the effect) is systematically tracked back to possible cause. The diagram organizes the search for the root cause of the problem. Figure 1 shows a sample cause and effect diagram.
Fig. 1 Sample cause and effect diagram
1.1.5 Histogram A bar graph shows the comparative frequency of a specific measurement. The shape of the histogram can indicate that a problem exists at a specific point in a process. 1.1.6 Control Charts A broken-line graph illustrates how a process on a point in a process behaves over time. Samples are periodically taken, checked, or measured, and the results plotted on the chart. The Lalit Kumar and D. R. Prajapati
chart can show how the specific measurement changes, how the variation in measurement changes, or how the proportion of defective pieces changes over time. Control charts are used to find sources of specification variation (variation that is caused by specific, fixable occurrences, to measure the extent of common cause variation (variation that is inherent to the process), and to maintain control of a process that is operating effectively. Figure 1.6 shows a sample of X-bar and R control chart.
Page 65
Proceedings of National Conference on Advancements and Futuristic Trends in Mechanical Engineering; of Mechanical Engineering, PEC University of Technology, Chandigarh on 17th-18th Oct. 2014 1.1.7 Scatter Plot Pairs of measurement are plotted on a two dimensional coordinate system to determine if a relationship exist between the measurements. Literature review on the applications of some important tools is presented in the following section.
2. Literature Review Shewhart [14] was the first, who suggested the X chart; using two control limits. Since then, various researchers did a lot of work on SPC techniques and control charts and it is discussed in this section. Deming [9] developed a new quality-improving concept, which extended quality inspection to statistical process control (SPC). The activities of off-line quality control are promoted by Design of Experiments (DOE).Various related research papers are summarized here in this section. Cheikh and McGoldrick [6] discussed the work, which has been carried out in the area of tolerances with cost, function and process capability the main parameters in mind. So that the resulting functional variables of the assembly can meet their respective functional tolerances requirement and cost of manufacturing all the components to their respective functional tolerances is minimized. It shows that, when manufacturing cost information and process capability information are available, functionally correct design at minimum cost can be achieved. Furthermore, they showed how statistical analyses of the manufacturing processes involved can lead to the relaxation in requirement at the same time, maintaining the desired levels of product quality and reliability. Wu [15] presented an approach to determine the optimum control limits of the x-bar chart for skewed process distributions. The approach takes both the control limits of the x-bar chart and the specification limits of x-bar into consideration, and relates the out-of-control status directly with the nonconforming products. The proposed approach may be applied to industries to reduce the average number of scrap products, without increasing the type I error in statistical process control (SPC).Chen and Ding [7] proposed a new index Spmk for any underlying distribution, which takes into account process variability, departure of process mean from the target value, and proportion of non-conformity. They first reviewed Cp, Cpk, Cpm and Cpmk, and their generalization, CNp, CNpK, CNpm and CNpmk. Then they proposed a new index S pmk. Proportion of nonconformity can be exactly reflected by S pmk. They demonstrated superiority of Spmk over CNpmk with several non-normal processes. CNpmk is a recently developed index which is related to process variability and departure from the target value. A method is proposed to estimate S pmk, with illustrations. MacCarthy and Wasusri [10] reviewed non-standard applications of SPC charts reported in the literature from the period 1989 to 2000, inclusive. Non-standard applications are analysed with respect to application domain, data sources used and control chart techniques employed. The principal application domain for statistical process control (SPC) charts has been for process control and improvement in manufacturing businesses. Balamurali and Kalyanasundaram [4] utilized a non-parametric but computer intensive method called Bootstrap. Usually the Lalit Kumar and D. R. Prajapati
Department
capability indices widely used by quality professionals as an estimate of process capability. Many process indices have been proposed and developed with Cp, Cpk and Cpm among the most widely used. More recently, techniques have been developed to construct lower 95 percent confidence limits for each index. These techniques are based on the assumption that the underlying process is normally distributed. The proposed method is defined for non-normal distribution. A simulation using three distributions (normal, log-normal and chi-squared) was conducted and a comparison was made of the performances of the Bootstrap and the parametric estimates. Casalino et al. [5] studied Ti6Al4V which presently is one of the most widely used titanium alloys, accounting for more than 50% of all titanium tonnage in the world, and to date no other titanium alloy has been a threat to its dominant position. Laser welding of Ti6Al4V is a major issue in the automotive and aerospace industries. In their study, both CO2 and diode laser welding processes were investigated for Ti6Al4V alloy sheet joining using either lap or butt configurations. Artificial neural networks (ANN) processed the data coming from the experimental trials. The aim was to interpolate the database in order to form a suitable database for the analysis of the variance (ANOVA) and the Taguchi analysis of the means. Anawa and Olabi [2] studied the most common problem of welding dissimilar metals (DMWs) with respect to residual stresses due to the differences in the coefficient of thermal expansion and heat conductivity of the two welded metals. In the work, a CO2 continuous laser welding process was successfully applied and optimized for joining a dissimilar AISI 316 stainless steel and low carbon steel plates. The Taguchi approach with three factors (selected welding parameters) at five levels each (L3-25) was applied to find out the optimum levels of welding speed, laser power and focal position for CO2 keyhole laser welding of dissimilar butt weld. Azzabi et al. [3] proposed a method to enforce six-sigma to assure high-level quality products, and to make firm a level of improvement for the long-term performance. The total performance of the process and the quality of it production depends on the one hand, of the characteristics of the intermediate products, and on the other hand, of the operation parameters of the manufacturing. To help accomplish this objective, various quality improvement philosophies have been put forward in recent years and of these, six-sigma has emerged as perhaps the most viable and efficient technique for process quality improvement. In this paper the application of six sigma methods enforces with multi criteria approach to permit classification the betters’ choices of a Tunisian industry. Prajapati and Mahapatra [11] discussed the limitations of CUSUM and EWMA charts and proposed a new chart and compared this chart with conventional CUSUM and EWMA charts. They found that CUSUM and EWMA schemes are ineffective to catch the process shifts when samples are not taken from same stream. And discovered that new (proposed) chart provides the highest percentage of correct signals compared to CUSUM and EWMA schemes for all process shifts. They also stated that the average run length (ARLs) of proposed scheme is much lower than both CUSUM and EWMA schemes for all the shifts in the process average. Abdolshah et al.[1] presented a review of loss-based PCIs such as Cpm, Cpmk, PCIθ, Cpc, Le and L″e. They also discussed the
Page 66
Proceedings of National Conference on Advancements and Futuristic Trends in Mechanical Engineering; of Mechanical Engineering, PEC University of Technology, Chandigarh on 17th-18th Oct. 2014 characteristics of loss-based PCIs such as reject based, asymmetric, bounded, loss based and target based. Finally they recommended development of a new loss-based process capability index with more excellent specifications. Das (2012) discussed the use of generalized lambda distribution to handle non-normal data. Traditional control chart has been established based on the assumption of normality. In many practical situation assumption of normality is violated. Under these situations, the use of traditional control chart gives erroneous conclusion. But for handling non-normal data one approach is use of non-parametric control charts which are not so efficient. Another approach is to use generalised distribution is very effective in non-normal data. Prajapati and Singh [13] provided a survey and brief summary of the work on the development of the control charts for variables to monitor the mean and dispersion for auto correlated data. Auto correlation is the stage when observation from many processes leads to automatically building-up correlation in the entire process. This autocorrelation among the observation can have significant effect on the performance of a control chart. The detection of special cause/s in the process may become very difficult in such situations. They found that various new methodologies and approaches such as double sampling, variable sample sizes and sampling intervals, etc. are suggested by various researchers to counter the effect of autocorrelation. Das and Sachan [8] discussed the importance of control charts in detecting the assignable cause of variation. They discussed the assumption under which these charts are
Wire Drawing
Wire
CNC turning 1st side
CNC turning 2nd side
developed. They proposed some alternatives control charts for controlling location parameters based on some robust estimators, because the present charts are not used with assumption in real situations. They also showed the performance of proposed control charts and compared them with some existed robust control chart.
3. Industry and Products This firm is one of the twelve collaborating industries and one of the largest firms of the fastener manufacturing. It is situated in the northern India. This company is established in 2011. This company supplies the products to: Hero Honda, Honda, Bajaj, Maruti-Suzuki, Yamaha, Bajaj, TVS and Mushashi etc. There are many products which are manufactured by this firm, like axles, bearing, nut and bolts, cams, brakes, drive shaft, kick gear starter, sprocket cam drive, axles, bearings etc. 3.1 Drive Shaft Manufacturing Process The drive shaft is used to transmit torque from engine. Drive shaft is an alternative to chain and belt drives and is used to distribute the torque from the transmission to the rear differential. Diameter ranges of drive shafts are: (i) 17.21 mm17.24 mm (ii) 16.82 mm-16.88 mm (iii) 15.42 mm-15.15.48 mm (iv) 12.21 mm – 12.24 mm (v) 11.35 mm- 11.355 mm (vi) 9.115 mm- 9 .125 mm. The manufacturing process of drive shaft is shown in Figure 2.
Cold forging
Final inspection
Normalising
packing
Dispatch
Fig. 2 Flow diagram of manufacturing of drive shaft 3.2 Pareto Diagram for Finding Major Problem in Micro-Turner-Vi The data for the month of February, 2014 is shown in the Table 1 Lalit Kumar and D. R. Prajapati
Department
Page 67
Proceedings of National Conference on Advancements and Futuristic Trends in Mechanical Engineering; of Mechanical Engineering, PEC University of Technology, Chandigarh on 17th-18th Oct. 2014 Table 1 percentage of defects of various components Sr.No.
Name of Defect
Frequency of defects
Cumulative frequency
Percentage
1
drive shaft run out
477
477
29.75%
2
Sprocket I.D. undersize
265
742
45.46%
3
255
997
61.09%
216
1213
74.32%
5
Gear Kick Starter Pitch Circle Diameter run out Gear kick starter drive outer diameter sprocket I.D Oversize
92
1305
79.96%
6
drive shaft length oversize
76
1381
84.62%
7
69
1450
88.85%
8
drive shaft length undersize G.K.S's I.D Undersize
67
1517
92.95%
9
Sprocket O.D oversize
47
1564
95.83%
10
sprocket Oval D
37
1601
98.1%
11
G.K.S's I.D oversize
18
1619
99.2%
12
Sprocket O.D undersize
13
1632
100.00%
4
The Pareto diagram for problem analysis in Micro-Turner-VI is shown in Fig. 2.
Fig. 3 Pareto diagram for problem analysis in Micro-Turner-VI
Lalit Kumar and D. R. Prajapati
Page 68
Department
Proceedings of National Conference on Advancements and Futuristic Trends in Mechanical Engineering; of Mechanical Engineering, PEC University of Technology, Chandigarh on 17th-18th Oct. 2014
Department
It is observed from the Pareto analyses that there are five main major defects which contribute maximum rejection level of the industry and these are
of control. If the processes will be out of control then causeand-effects analysis will be done to find the root causes of the problems. Control charts are applied in the mains products, having higher defect rate and discussed in the following section.
1. 2. 3.
3.4 Application of Cause and Effect Diagram to Find the Root Causes
3.3 Results of Pareto Analysis
4. 5.
Drive shaft Run-Out (29.75%). Sprocket Inner-Diameter undersize (16.24%). Gear Kick Starter Pitch Circle Diameter Run-Out (15.63%) Gear kick starter drive outer diameter (13.24%) Sprocket Inner-Diameter Oversize (5.64%).
All above five defects contribute to more that 75% of the total defects occurring in the industry. With the help of Pareto analysis main defects affecting the company are identified. First four out of five problems are taken for analyses. Control charts will be used to check whether the processes by which these products are being manufactured are under-control or out
A Cause-and-effect diagram is a tool that helps to identify, sort, and display possible causes of a specific problem or quality characteristic. It graphically illustrates the relationship between a given outcome and all the factors that influence the outcome. This type of diagram is sometimes called an "Ishikawa diagram" because it was invented by Kaoru Ishikawa, or a "fishbone diagram" because of the way it looks. The structure provided by the diagram helps the team members to think in a very systematic way. Figure 3 shows cause and effect diagram for drive shaft.
Fig. 4 Cause and effect diagram for drive shaft run-outs 3.5 Recommendations to Minimize the Problem of RunOuts From the Cause and effect diagram; shown in Figure 3, following suggestions may be incorporated to minimize the drive shaft run outs: (i) In-process inspection must be ensured for each manufacturing operation. Lalit Kumar and D. R. Prajapati
(ii) Cutting speed should be optimum. (iii) Inspection should be done meticulously and carefully. (iv) Insert should be replaced before it becoming blunt. (v) Jaw and chucks should be installed carefully. (vi) Revolving centre should be replaced periodically.
Page 69
Proceedings of National Conference on Advancements and Futuristic Trends in Mechanical Engineering; of Mechanical Engineering, PEC University of Technology, Chandigarh on 17th-18th Oct. 2014
Conclusions Many companies that initially aimed at improving the quality of their products found that to satisfy the final customer. It was necessary to satisfy a whole sequence of internal customers. Statistical process control is an industry-based methodology for measuring and controlling quality during the manufacturing process. Quality data in the form of product or process measurement are obtained in real-time during manufacturing. In quality control, it often represents the most common sources of defects, the highest occurring type of defect, or the most frequent reasons for customer complaints, and so on. Pareto charts for the third and fourth quarters of the current year have been plotted and major causes of rejection have been explored. It is concluded that major defects occur in drive shaft, sprocket inner-diameter, gear kick starter PCD and gear kick starter drive outer diameter.
References 1.
2.
3.
Azzabi L., ayadi D., boujelbenne Y., Kobi A., Robledo C. & Chabchoub H., 2009, Six sigma based multicriteria approach to improve decision setting, Inteternational journal of quality engineering and technology, Vol. 1, No. 1, pp.99-123.
4.
Balamurali S. & Kalyanasundaram M., 2002, Bootstrap lower confidence limits for the process capability indices C, C and C, International Journal of Quality & Reliability Management, Vol. 19 Iss: 8/9, pp.1088 – 1097.
5.
6.
Cheikh, A. and McGoldrick, P.F. (1988), “The Influence of Cost, Function and Process Capability on Tolerance”, International Journal of Quality & Reliability Management, Vol. 5, Issue 3, pp.15 – 28.
7.
Chen, J.P. and Ding C.G. (2001), “A new process capability index for non-normal distributions”, International journal of Quality & Reliability Management, Vol. 18 Issue 7, pp. 762 – 770.
8.
Das, N. (2012), “Performance of control chart using generalised lambda distribution, International journal of Productivity and Quality Management, 2012 Vol.10, No.4, pp.411 – 427.
9.
Deming, W. E. (1982), “Out of the Crisis”, Unlimited Learning Resource, LLC, Winston-Salem, North Carolina.
10.
MacCarthy, B.L and Wasusri, T. (2002), “A review of non-standard applications of statistical process controls (SPC) charts”, International Journal of Quality & Reliability Management, Vol. 19, Issue 3, pp. 295 – 320.
11.
Prajapati, D.R. and Mahapatra, P.B. (2009), “A new X chart comparable to CUSUM and EWMA charts, International journal of Productivity and Quality Management, Vol.4, No.1, pp.103 -128.
12.
Prajapati D.R. & Singh S., 2012, Control charts for monitoring the autocorrelated process parameters: a literature review, International journal of Productivity and Quality Management, 2012 Vol.10, No.2, pp.207 – 249.
13.
Das N., & Sachan L., 2013, Robust control charts for controlling location parameter, International journal of Productivity and Quality Management, 2013 Vol.12, No.1, pp.18 – 37.
14.
Shewhart, W. A. (1931), “Economic control chart of quality of manufacturing product”, Van Nostrand, New York.
15.
Wu, Z. (1996), “Asymmetric control limits of the -bar chart for skewed process distributions”, International Journal of Quality & Reliability Management, Vol. 13 Issue 9, pp.49–60.
Abdolshah, M., Yusuff, R.M., Hong, T.S. and Ismail, M.Y. (2011), “Loss-based process capability indices: a review”, International Journal of Productivity and Quality Management, Vol. 7, No.1, pp. 1 – 21. Anawa, E.M. and Olabi, A.G. (2008), “Using Taguchi method to optimize welding pool of dissimilar laser-welded components” Original Research Article Optics & Laser Technology, Volume 40, Issue 2, March 2008, Pages 379-388.
Casalino, G., Curcio, F. and Minutolo, F.M.C. (2005) “Investigation on Ti6Al4V laser welding using statistical and Taguchi approaches.” Journal of Materials Processing Technology, Volume 167, Issues 2–3, pages 422-428. .
Lalit Kumar and D. R. Prajapati
Department
Page 70
