joining of gfrpin marine applications

In: Advances in Materials Science Research. Volume 12 ISBN 978-1-62100-045-7 Editor: Maryann C. Wythers © 2012 Nova Science Publishers, Inc.

Chapter 6

JOINING OF GFRP IN MARINE APPLICATIONS Authors* Di Bella Guido - CNR ITAE Galtieri Giovanna - Department of “Chimica Industriale e Ingegneria dei Materiali”, University of Messina Pollicino Enzo - Department of “Chimica Industriale e Ingegneria dei Materiali”, University of Messina Borsellino Chiara - Department of “Ingegneria Civile”, University of Messina

Abstract In recent years the introduction of glass fibre reinforced plastic (GFRP) in the recreational boat industry has been undoubtedly significant. Wood has been gradually replaced, thus modifying the classic canon of shipbuilding industry and leading to an industrial revolution that has strongly changed the way boats are designed and produced. GFRPs are widely used in the marine industry thanks to their good environmental resistance, the possibility of realising complex shapes and also their high specific strength and stiffness. The key feature is their weight, which has made the sandwich concept an attractive alternative to the traditional design concept. In fact, the growing request of larger and faster ships has increased the demand for lighter and stronger structures and for a better utilisation of both the materials and the involved structure. A lighter ship building allows to obtain: high acceleration, low consumption and, as a consequence, higher attention for the environment, possibility to use less powerful engines, higher cargos. A ship, made with GFRP, consists mainly of the following parts: - hull: frame or body of the boat. This structure is reinforced by internal transversal panels, i.e. multiple watertight bulkheads are used to divide the hull into many compartments. Such structural solution optimises the resistance to axial deformations, shear loads and unsymmetrical static or dynamic loads that can occur locally under working conditions; - deck: permanent covering over compartments or hull. On a ship, the primary deck is the horizontal structure which forms the 'roof' for the hull. It strengthens the hull and serves as the primary working surface as well; *

E-mail address: [email protected]

Di Bella Guido, Galtieri Giovanna, Pollicino Enzo, Borsellino Chiara -

superstructures: upward extension of an existing structure above a baseline. They are the parts of a boat that project above her main deck. All these parts and structural elements have to be connected. In particular, it is possible to identify three different joining systems: - over-lamination: between the parts, that have to be joined, several glass fibre fabrics (i.e. Mat) with polyester or vinyl-ester resin are laminated through a hand lay-up process; - adhesive bonding: a mixture in a liquid or semi-liquid state adheres or bonds items together. Adhesives may come from either natural or synthetic sources. Some modern adhesives are extremely strong, and are becoming increasingly important in modern construction and industry; - mechanical fastener: device that holds two or more objects together. The structures are drilled and then assembled by a fastener (i.e. a button or a zipper as well as a bolt or a screw). The aim of this work is to give an outlook of the characteristics of these kinds of GFRP structures‟ joining in terms of advantages and limits, failure modes, configurations, applications, finite element modelling.

1. Introduction In the naval industry the need for lighter and faster boats, and, recently, boats with low environmental impact, has given more importance to the fibre reinforced plastic (FRP) materials or composites, reducing the use of metals such as steel and aluminium. A typical composite contains layers of multi-axial fibreglass fabrics reinforced with resin matrix, giving high strength and stiffness in different directions. Composites also improved corrosion resistance, fuel efficiency and reduced magnetic signature. The ship design consists of three principal parts:  Hull: frame or body of the boat, where it is possible to distinguish among the keel, the single or double bottom and sandwich laminates for side shell. The inner hull is stiffened by longitudinal and transverse stiffeners (longitudinal with web-frames, floors, beams or frames with stringers, girders or keelsons).The hull is divided into many compartments through internal transversal panels, i.e. multiple watertight bulkheads. Such structural solution optimises the resistance to axial deformations, shear loads and unsymmetrical static or dynamic loads that can occur locally under working conditions;  Deck: permanent covering over compartments or hull. On a ship, the primary deck is the horizontal structure which forms the 'roof' for the hull. It both strengthens the hull and serves as the primary working surface. The main purpose of the primary deck is structural, but it also provides weather-tightness, and supports people and equipment. The deck functions as lid for the complex box girder of the hull. It resists tension, compression, and racking forces.  Superstructures: upward extension of an existing structure above a baseline. They are a decked structure located above the weather deck (the weather deck is the uppermost complete weathertight deck fitted as an integral part of the yacht 's structure and which is exposed to the sea and the weather), extending from side to side of the hull or with the side plating not inboard of the shell plating more than 4% of the local breadth. Superstructures may be of different tiers in relation to their position in respect of the weather deck. A 1st tier superstructure is one fitted on the weather deck, a 2nd tier superstructure is one fitted on the 1st tier superstructure, and so on.

Joining of GFRP in marine applications

187

In every product, several parts or components are joined together to make the complete assembly. For example, there are several thousands of parts in an automobile, a yacht, or an aircraft. The steering system of an automobile has more than 100 parts. Heloval 43-meter luxury yacht from CMN Shipyards is made up of about 9000 metallic parts for hull and superstructure, and over 5000 different types of parts for outfitting. These parts are interconnected with each other, thus originating the final product. The purpose of the joint is to transfer loads from one member to another, or to create relative motion between two members. A good design policy suggests to avoid joints, as they are the weakest areas and most failures arise from joints. Therefore, joints are eliminated by integrating the structure. Joining has the following disadvantages: 1. it is a source of stress concentration. It creates discontinuity in the load transfer; 2. the creation of a joint is a labour-intensive process; a special procedure is followed to make the joint; 3. it adds manufacturing time and cost to the structure. In an ideal product, there is only one part. Fibre-reinforced composites provide the opportunity to create large, complicated parts in one shot, contributing to the reduction of the number of parts in the whole structure. There are three types of joints used in the fabrication of composite products, in naval applications: - Over-lamination; - Adhesive bonding; - Mechanical joints. Over-lamination is actually employed to connect several parts by a further handlamination. In adhesive bonding, two substrate materials are joined by an adhesive. Mechanical joints for composites are similar to the mechanical joints of metals. In this case, rivets, bolts, and/or screws are used to form the joint.

Di Bella Guido, Galtieri Giovanna, Pollicino Enzo, Borsellino Chiara

2. Overlamination Overlamination is used to connect bulkheads and other important reinforcing elements to the adjacent structure on both sides, where possible. It consists of overlapping the layers made of fibreglass reinforced resin to form a large patch fitted for the corner profile, known as T-joint o corner joint.

GAP

Figure 1. An example of over-lamination sequence. It is important that the taper of the tissues is set as shown in the Figure 1, so with each layer of at least 25 mm longer than the previous (if not every layer would act only on the previous layer and not on the surface of the elements to be connected) with the exception of the first that must overlap by at least 50mm on both the connecting elements. Then the edges of the reinforcements of one layer must not only be juxtaposed, instead it is necessary to overlap with the same offset various successive layers [1]. A critical area to take into account is the zone under the first layer (gap) that usually is not able to fill correctly the corner between the joining elements (see Figure 1)

Figure 2. T-joints between deck and bulkhead.


189

Figure 3. Deck-side shell connection. Figure 2 and Figure 3 show some examples of overlamination. A bulkhead - deck (or hull) connection (i.e. T-joint) is needed in order to distribute flexural, shear and tensile loads between the structures linking. Usually, the layers of overlamination are made from reinforcements by chopping fibre reinforced with resin (Mat), but it is also possible to use multiaxial fiberglass fabrics (bidirectional, quadriaxial or combi-fibers). The choice of the over-lamination orientation does not change the mechanical properties of the T-joint. The bulkheads are made in glass-reinforced plastic (GRP) or a wood based sandwiches (i.e. with balsa core). Di Bella et al. [2] have worked on adhesive bonded joints on ship structures; particularly T-joints with three configuration: the first was a methacrylate adhesive joint and the other two were over-laminations with different fibre‟s orientation (i.e. stratification layers made of 0/90 and +45/_45 bidirectional glass fabrics). It was found that when the elements that constitute the T-joints are made in GRP composite, the over-lamination enhances the mechanical properties in relation to the adhesive joint. For the wood sandwich, no difference between the use of the over-lamination and structural adhesive was evidenced. Moreover, it is possible to affirm that, for the typical observed failure mechanisms, the orientation of the over-lamination does not vary the behaviour of the joint. The deck - side shell connection has to be designed focusing on both the bending stress, caused by vertical loads on deck and horizontal loads of seawater, and the shear stress caused

Di Bella Guido, Galtieri Giovanna, Pollicino Enzo, Borsellino Chiara by the longitudinal bending. For such purpose a PU (Polyurethane) reinforce is used (see Figure 3). Another example of overlamination is the connection of primary stiffeners with the hull to ensure structural continuity. The type of connection can be “double zeta” or “omega”. The first, Figure 4a, involves overlapping layers from the right side with those from left by interposing some additional unidirectional layers. In the second, Figure 4b, the overlapping layers follow the shape of the structure extending the stratification at least 100mm on the other element which has to be joined. In both cases unidirectional fiberglass fabrics, arranged in the longitudinal direction, and a filler to fill the corner gap are added.

a)

b) Figure 4. Types of overlamination: a) double Z; b) omega. Figure 5 shows the crossing between the intersection of two stiffeners on the bottom hull, which should be watertight in order to ensure a perfect insulation of the core (PVC or PU) structures. Beforehand, it is necessary to laminate the quarter FRP tube both at the bottom hull and at one stiffener by means of layers stratification with adequate taper, and afterwards the other structure, previously hole-shaped, can be placed.


191

Transverse stiffener

¼ FRP Tube Overlamination

Figure 5. The crossing stiffeners. It is possible to find passages trough the hull, especially at the level of the rudder stock, shaft brackets, or bulkheads. Such crossings need to be built strongly; in particular, if subjected to alternating loads, they should also be reinforced by means of a plate and counterplate connected to each other with overlamination. Moreover, passages through watertight bulkheads are produced with the tapered stratification to be watertight as the bulkheads. At the intersection of two stifferners the lamination of upper surfaces must be continuous throughout the stiffener‟s length and it has to be performed by overlapping the unidirectional fibreglass for not less than 300mm from the cross. The connection between structural elements has to be made by corner joint (Figure 6).

Figure 6. The crossing stiffeners. The critical point in the corner joint is the gap, as shown in Figure 1. Between the first and second layer of overlaminate is advisable to place some fibre roving around each corner, to ease the subsequent application of tissue and to prevent the resin accumulation. The

Di Bella Guido, Galtieri Giovanna, Pollicino Enzo, Borsellino Chiara process continues by overlaying all the layers in accordance with the approved plan of lamination, making sure to thoroughly wet the fabric reinforcement. When using hand lay up technique, it is necessary to use always the "bubble breaker" in order to eliminate any air bubbles that would otherwise remain trapped in the laminate, push out resin in excess from the laminate, thereby obtaining the optimum fiber/resin ratio. Usually the gap is filled with a structural adhesive, like a filler, sometimes reinforced with fiberglass.


193

3. Adhesive joining If an effective bonding technology with a relatively high strength/weight ratio and reduced cost is needed, the adhesive joining is the most common, especially for marine applications. Such trend is particularly clear in the field of boats construction. Their simpleness and easiness of application, together with their elastic properties after curing, are the peculiarities that make adhesive joinings ideal for the construction of pleasure boats, yachts, ferries, cruise ships and even offshore boats. Bonding is undoubtedly the most universal joining technique. For what concerns ships, this techniques allows to avoid the discontinuities due to mechanical bolts, and, at the same time, to realize complex joint geometries and to homogenously distribute stresses on the whole joined surface, as to dampen vibrations and to obtain lighter structures because of the absence of bolts or rivets. Along with this advantage, the adhesive provides a layer between the materials that can absorb energy in impact, allowing for thermal expansion and acting as a layer which increases resistance to fatigue. The above reported benefits consent to attain lower production costs and/or better quality joints, when adhesives are used complementarily with traditional connection techniques in several application.

3.1. Adhesives Adhesives can be classified according to their application method. There are: pressure sensitive adhesives, hot models adhesives, chemical activated adhesives, and so on. For hot models adhesives, in particular, if the temperature required to establish the bonding is considered a key feature, the adhesives can be classified according to T[°C]. Therefore the adhesives will have an activation: - at temperatures below 20° C (cold); - at environment temperature, between 20 and 30° C; - at intermediate temperature, between 30 and 100° C; - at temperatures above 100° C (hot). Finally, another distinction is done between structural and non-structural adhesives. This classification is partially arbitrary, as there is no specific definition of "structural". The term “structural adhesive” typically defines an adhesive which can withstand high loads under a variety of different stresses for long periods of time. The non-structural adhesives are mainly used to create temporary adhesion. A conventionally criterion for defining a structural adhesive can be considered the shear stress: it must be above 10 N/mm2 at room temperature. Notwithstanding the foregoing, it is usually very difficult to rank the adhesives according to a single parameter, because some constituents may have properties of two different categories, as they might have been obtained by mixing constituents belonging to different categories. Generally, the widest used classification is that one based on the division of adhesives in three categories: - thermosetting; - thermoplastic; - elastomers.


The thermosetting resins are organic compounds that harden after polymerization. At high temperatures, approximately 250°C, they do not liquefy, instead they undergo a process of degradation and physical properties decay. Their mechanical properties are characterised by good shear strength and creep resistance, but low peeling resistance. The mentioned features limit the use of thermosetting resins in heavy load condition; they are mainly used as bonder between metals, wood parts or FRP composites. On the converse, the thermoplastic resins cannot be used at high temperatures, as they become soft, and they have good peel resistance. Elastomers instead are widely used to modify thermosetting system. An elastomer can usually stretch twice than its original length without experiencing a decrease in its elastic properties. In addition to this, they have high peel strength and flexibility, but low shear strength and poor creep resistance. Their max operating temperature is between 80 and 100°C, while the silicone rubber reaches 200°C [3]. The most common adhesive families used, as structural adhesives, are: - Acrylic; - Anaerobic; - Cyanoacrylate; - Epoxy; - Hot Melt; - Methacrylate; - Polyurethane; - Silicones. Acrylic Adhesives have formulations that tolerate dirtier and less prepared surfaces generally associated with metals. They challenge epoxies in shear strength, and offer flexible bonds with good peel and impact resistance. Acrylics are two-part adhesives, the resin is applied to one surface and an accelerator or primer to the other. The two parts can be preapplied and later mated. Once mated, handling strength is typically achieved in a few minutes. Curing can be completed at room temperature. Newer versions of acrylics are now available in two component formulations that are mixed together before the application. Anaerobic adhesives are among the most common structural adhesives. As the curing mechanism is triggered by deprivation of oxygen (hence the name „anaerobic‟ or „without air‟), anaerobic adhesives will not cure prematurely. These adhesives are based on acrylic polyester resins and are produced in viscosities ranging from thin liquids to viscous thixotropic pastes. Even though they have high cohesive strength, they have low adhesive strength and do not fit to permeable materials. Anaerobics do not fill gaps well and may require primers. They are generally used as thread fasteners. Cyanoacrylate Adhesives (superglues) are also easily applied and offer extremely fast cure rates. Cyanoacrylates are relatively low viscosity fluids based on acrylic monomers and, when placed between closely fitting surfaces, some will cure to a strong joint in two or three seconds. The capacity of cyanoacrylates to join plastics and rubbers to themselves or to other substrates is their biggest advantage. On the other hand, cyanoacrylate adhesives exhibit poor impact resistance, are vulnerable to moisture and solvents, and are only suitable for small areas. In addition, they do not fill gaps well, require precise mating of bonded surfaces, and are relatively expensive. They also have poor solvent and water resistance. Epoxy Adhesives have been available longer than any engineering adhesive and are the most widely used structural adhesive. Epoxy adhesives are thermosetting resins which solidify by polymerisation and, once set, will soften but not melt on heating. Two part resin/hardener systems will solidify on mixing (sometimes accelerated by heat), while one


195

part materials require heat to initiate the reaction of a latent catalyst. Epoxies offer very high shear strengths, and can be modified to meet a variety of bonding needs. Generally epoxy bonds are rigid: they fill small gaps well with little shrinkage. Hot Melt Adhesives have moved out of their traditional applications into areas of lowstress product assemblies. They form flexible and rigid bonds, achieve 80% of their bond strength within seconds, bond permeable and impermeable materials, and usually require no elaborate surface preparation. Hot melts are insensitive to moisture and many solvents, but they soften at high temperatures. Methacrylate Adhesives provide a unique balance of high tensile, shear and peel strengths with the maximum resistance to shock, stress and impact across a wide temperature range. Methacrylates can generally be used without surface preparation when joining plastics, metals and composites. They are two component reactive materials based on methyl methacrylate monomer that, when mixed together, have a controlled cure speed based on the appropriate application process. Methacrylates are tolerant to off ratio mixing and remain strong and durable under severe environmental conditions. They resist water and solvents thus forming an impenetrable bond. Methacrylate adhesives, belonging to structural adhesives, have been developed to perform under a wide range of conditions. This results is achieved by improving curing, bonding, adhesion and toughening mechanisms. They have a unique combination of polymers and impact modifiers which enable them to be strong and yet flexible. The resulting high tensile strength along with impressive elongation (over 100% at 30°C) gives them the processing advantages and convenience of epoxies and polyurethanes, without increased cost. The total combination of all these factors along with good rheology, thixotropic properties, ease of use and lack of VOC‟s and solvents, results in a range of structural methacrylate adhesives that can be successfully used in the marine markets on a wide range of substrates. The real revolution is in the bonding of deck to hull and stringer bonding. Not only does bonding with methacrylate adhesives save time and reduce labour costs but it also provides a strong, durable, flexible, energy absorbing bond line, which is not susceptible to cracking or brittleness. One of its application can be the bonding of the hull to the deck or the bulkheads to the side shell of hull. The interfaces between the adhesive and the substrate are critical areas because of the difference of stiffness and the stress concentration effect near the overlapping between the two different materials. Figure 7 shows T-joints bonding with methacrylate adhesive simulating bulkhead to hull connection undergone to tensile stress. The joint a) is characterised at first by the shear collapse of the plating sandwich and then by the failure at the interface between the base and the bulkhead and between the adhesive and the bulkhead. The fracture always occurs at the substrate of the bulkhead. The joint b) is characterised firstly by shear cracks propagation in the plating sandwich and then by the failure of the joint. The fracture occurs at the substrate of the base evidencing good adhesion [2]. Polyurethane Adhesives are named after the polymer type formed on completion of the reaction. The adhesives are made up of two component: one side is always based on isocyanate, while the other might be made with several core reactants often amines or glycols. They are known for toughness and flexibility even at low temperatures. They have fairly good shear strength and excellent water and humidity resistance, although uncured urethanes are sensitive to moisture and temperature [4]. Silicone Adhesives. The Silicone is a semiorganic polymer. It may be fluid, gel, elastomeric, or rigid in their form. All silicone polymers share the following features: - low surface tension; - high lubricity with rubber and plastic surface;


excellent water repellency; good electrical properties; thermal stability; chemical inertness; resistance to oxidation, hot water and weather; flexibility at also low temperature; excellent temperature resistance from -200 to 260°C.

a)

b) Figure 7. Failure mechanisms of methacrylate adhesive in T-joint between: a) GRP sandwich (plating element) and wood sandwich (i.e. balsa core); b) GRP sandwiches. As the latest generation of polymers, “MS” polymer comes from the polyurethanes with enhanced properties. The MS abbreviation stands for “sealant modified”. They are good UV resistance; they dampen the vibrations and they show good mechanic properties also as adhesive. They can be painted and sanded and they do not contain isocyanates; they exhibit short drying times, good adhesion on various substrates. They are recommended for important structural bonding operations and are ideal for external use such as in the teak, porthole and hull sealing operations. In modern boats, timber decking is frequently built in the form of prefabricated panels laid over the structural deck. These panels generally consist of a marine grade bonded plywood backing with strips of teak or Oregon pine bonded or glued to the face. Another type of prefabricated panel consist of teak planking with rubber jointing strips and no plywood backing. The polyurethane adhesives are ideal for bonding these panels to the deck, because they are resistant to sea water, they possess excellent gap filling properties, and no additional mechanical fastenings are needed. Post cured, the adhesive bond is strong, permanently elastic (in the summer weather the joint expands and contracts), flexible and possess good gasket or sealing properties. The curing occurs thanks to the adsorption of moisture from the air or from the joining surfaces so it is rather slow. For their waterproof qualities the adhesives can be applied to the whole surface on the deck as an additional protective skin from the attack by natural elements.


197

After prolonged periods of exposure the joint surface may exhibit slight degradation, but this does not detract from the sealing properties as it is a only surface effect. UV-resistant polyurethane adhesives were formulated to make a durable surface of joint against solar UV radiation and sea water attack. These adhesives are used in the glazing of mineral or plastic (PMMA and PC) glasses into frames or directly into the hull or deck. When designing glazing installations an expansion gap must be incorporated between the window rebate and the glazing panel to accommodate thermal movement, as plastic glazing products have a higher coefficient of thermal expansion than conventional glass. To minimise the risk of stress cracking, flat sheet of glazing material should be completely flat; they should not be forced to take up a curvature. After the bonding the sealing must be protected against UV radiation and water penetration [5].

3.2. Surface Pre-treatments The adhesive bonding behaviour depends on many factors such as physical and mechanical properties of both adherent and adhesive, the geometrical parameters of the joints like the overlap length, the overall length of sample, the thickness of the substrates and adhesive, the surface finish. Surface preparation is an important process step governing the quality of an adhesive bond. The pre-treatments increases the bond strength by altering the substrate surface in a number of ways, i.e. increasing surface tension and roughness or changing surface chemistry. The nature of the surface will also influence the stability of the joint. When exposed to hot/humid environmental condition, a polymeric adhesive/polymer interface is much more stable than the equivalent polymeric adhesive/metal interface. The efficiency of a surface treatment depends on the nature of the substrate and on the depth of the treatment. A compromise between the functionality and the degradation of the surface might need to be reached. There are many available methods to pre-treat polymers and metallic alloy substrates. More specifically, they can be chemical or electrochemical, mechanical, thermal, photochemical, or plasma. With chemical and electrochemical methods a number of changes may occur, namely, removal of weak material (i.e. material with low cohesive strength is removed and the topography may be changed), roughening, and the introduction of functional groups into the polymer. It is possible to have acid and alkaline etching. The first produces similar results to abrasion and grit blasting, showing an increase in bond strength for thermoset polymer composites, whereas little or no effect was observed on thermoplastics. The substrate, previously degreased, has to be dipped into an acid solution in estimated temperature and time condition. An additional problem associated with these treatments is hydrogen pick up. Nevertheless, this problem can be avoided by using alkaline solutions instead of acid etchants. In alkaline solutions, depending upon the concentration of sodium hydroxide and hydrogen peroxide, the metal is either etched or oxidised. Those concentrations, which produce grey oxides, have been found to produce adhesive wetting surfaces. For example, in the case of aluminum and its alloys, for durable adhesively bonded joints a simple degreasing or grit blasting is insufficient, while electrochemical etching may be applied, usually the metal acts as an anode or cathode in an acid or alkaline solution [6]. In the marine industries these pre-treatment are not often applied, instead mechanical abrasion using abrasive papers is more common. The choice of this treatment is due to less waste of material, less costs in tooling and machinery, lower manufacturing times. The roughness, obtained with mechanical abrasion, leads to an increased contact area between the two substrates and increases the adhesion by mechanical interlocking [7]. To

Di Bella Guido, Galtieri Giovanna, Pollicino Enzo, Borsellino Chiara remove the residual particles remaining after mechanical abrasion is need to clean with pressurized air or solvents. In the contact adhesive-adherends the presence of air bubbles inside the adhesive or entrapped between the substrate and adhesive, can lead to the formation of cracks as a consequence to cause an adhesion or cohesive failure (see Figure 8). s u b s t r a t e

adhesive

s u b s t r a t e

a

b

c

Figure 8. Failure modes of single lap joint: a) adhesive failure; b) cohesive failure; c) substrate failure The industrial use of adhesive joints with a good surface preparation takes into account the adhesive‟s costs such as to remain competitive when compared with other joint processes such as welding or riveting. Sometimes it is possible to achieve high performance of the adhesive joint with more economical and environmentally friendly surface finishes, i.e. abrasion with granulated emery paper rather than sand blasting [8]. Correct surface preparation is the key for a successful bonding. The level of cleaning will be determined by the type of surface and the degree of deterioration. If compressed air is used to remove dust from surfaces, the air should be filtered to remove traces of oil. The use of vacuum cleaners is even better for dust removal. Solvents, such as alcohol, are not recommended as they can hinder cure or subsequent adhesion. Always use clean and lint free wipes, and change them frequently to ensure that the contamination is removed from and not redistributed onto the surface. Once clean, the substrates should undergo a process of complete dryness before proceeding with the next operation. Priming is a means of transforming a surface, either chemically or physically, into an ideal condition for successful bonding thus ensuring long-term performance. The simplest form of priming is wiping the prepared surface with an activator which reacts with the surface providing improved “wetting” characteristics and more reactive sites. Porous and rough surfaces require a primer with film-forming properties to redefine the surface producing a denser more even bond line. Primers must always be allowed to dry thoroughly before application of the adhesive. If left too long, primed areas must be reapplied or reactivated. Primed surfaces should be protected from contamination by dust, dirt, grease, vapour, moisture, etc., until the bond is formed. For an example in the process of caulking teak deck priming the plank is a important step, each failure might be harmful to the final quality of the seal and will impair the longevity of the deck [5].


199

3.3. Finite Element Analysis Numerical analyses are a good tool to design a joint. In fact, they allow to investigate the mechanical behaviour. Moreover, coupled with specific failure criteria, they allow to predict the fracture. In Figure 9 the shear stress map for a fibreglass/polyester resin composites joined with an unsaturated acrylic polyurethane adhesive is reported, showing that the shear stresses are mainly distributed along the overlap length. The maximum values are in the edges. These local stresses are obviously higher than the mean experimental ones. The laminates do not support relevant shear stresses. To better understand this behaviour a quantitative analysis can be carried out. Particularly, Figure 10 shows the shear stress on the joint overlap length, at the substrates interface (left, right) and within the adhesive (centre) [9].

Figure 9. Shear stress in a single lap joint.


Figure 10. Shear stress in the overlap length of the joint. Figure 11 shows the distribution of the stresses along the direction of the y-axis for a Tjoint. The maximum value (i.e. identified by „„MX‟‟) is located at the corner where there is a stress concentration. This effect is reduced in order to reach a real value of stress by eliminating the shape corner in model design. In particular, a connecting line is introduced. The optimum line is curved, but in the mesh procedure, the corner elements are characterised by an area too small and as a consequence the simulation fails. To avoid this problem, a triangular loop line is modelled. In Figure 11, two zones are evidenced: a first zone subjected to a tensile stress and a second one to a compression stress. The separation line between these regions fits well the failure line, as evidenced in the photo of the fractured joint (see Figure 11 right). The separation line is at 0 stress and it corresponds to the failure location because it represents the interface between two zones characterised by stresses with opposite directions. As a consequence, a shear solicitation is produced. It is also possible to identify the point where the crack starts [2].


Figure 11. Distribution of the stresses along the direction of the y-axis.

201


4. Mechanical joining Mechanical joining is most widely used in joining metal components. Examples of mechanical joints are bolting, riveting, screw, and pin joints. Similarly to metal components, composite components are also joined using metallic bolts, pins, and screws, especially in marine applications. For most mechanical joints, an overlap is required in two mating members and a hole is created at the overlap so that bolts or rivets can be inserted. When screws are used for fastening purposes, mostly metal inserts are used in the composites, the reason being that the threads created in the composites are not strong in shear and therefore metal inserts are used. Figure 12 shows examples of bolting.

Figure 12. Bolted joint. In bolted joints, nuts, blots, and washers are used to create the joint. In riveting, metal rivets are used. Bolted joints can be a single lap joints, double lap joints, or butt joints, as shown in Figure 13.


203

Figure 13. Types of bolted joints.

4.1. Advantages and limits of mechanical joints The advantages of mechanical joints are: 1. They allow repeated assembly and disassembly for repairs and maintenance without destroying the parent materials; 2. they offer easy inspection and quality control; 3. they require little or no surface preparation. Whereas, the limits are: 1. Mechanical joints add weight to the structure and thus minimize the weight-saving potential of composite structures; 2. they create stress concentration because of the presence of holes. Composite materials, unlike ductile materials such as aluminium and steel, do not have the capacity to redistribute local high stresses by yielding. In composites, stress relief does not occur because the composites are elastic to failure; 3. they create potential galvanic corrosion problems because of the presence of dissimilar materials. For example, aluminium or steel fasteners do not work well with carbon/epoxy composites. To avoid galvanic corrosion, either metal fasteners are coated with nonconductive materials such as a polymer or composite fasteners are used; 4. they create fibre discontinuity at the location where a hole is drilled. They also expose fibres to chemicals and other environments.

4.2. Failure modes A bolted joint is made by drilling holes in mating parts. The mating parts are then aligned and a nut is passed through it and then bolted. Failure in a bolted joint may be caused by [10]: 1. Shearing of the substrate (shear out); 2. Tensile failure of the substrate (net tension); 3. Crushing failure of the substrate (bearing);

Di Bella Guido, Galtieri Giovanna, Pollicino Enzo, Borsellino Chiara 4. Shearing of the bolt. The net tension and shear out failure modes are catastrophic and contribute in different ways to fracture, for a given thickness, at varying distances between the hole centre and the free-end edges of the laminate [11]. In particular, shear out failure mode occurs mainly in joints, where the distance between the hole edge and the edge of the laminate is small, or in highly orthotropic laminates such as cross-ply laminates. Net tension failure, in tension or compression, can occur when the area of cross section of the sample is particularly small. Bearing failure is primarily a compressive failure occurring close to the contact region at the hole edge. The failure is caused by compressive stresses acting on the hole boundary [12]. Bearing failure is, like all compressive failures in a composite laminate, associated with delamination and ply buckling. This implies that the bearing strength is strongly affected by the lateral constraint of the material surrounding the loaded hole [13]-[14]. The bearing strength is also strongly dependent on the lay-up of the laminate [15] and it is sensitive to environmental conditions such as high temperature and high moisture content [16]. Bolt failures are not common because steel fasteners are very strong in shear. Moreover, in several laminates it is possible to observe a cleavage failure mode, which occurs when the fibres are oriented in load direction and it can be considered as a mixed failure between the shear out and net tension. Figure 14 shows these typical failure modes.

Figure 14. Failure modes.


205

4.2.1. Bearing In order to evaluate the bearing resistance, it is needed to investigate the joint carrying out two kinds of analyses: i) Bearing Compression Test; ii) Tensile Bearing Test.

4.2.1.1. Tensile Bearing Test Tensile bearing tests are performed (according to [17]) on rectangular samples with length L = 100 mm, width w = 15 mm, and thickness t = 4.6 mm. On each sample a hole is created, of a diameter D, which is 1–5 times greater than d, at a distance from the free edge E= 17.5 mm (see Figure 15).

Figure 15. Sample geometry. The holes are realized by milling and mechanical tests are performed by employing an Universal Testing Machine, with a speed of 1mm/min. Different tests can be performed varying geometrical characteristics (i.e. e/d and w/d ratios).

4.2.1.1.1 Prediction of failure modes Bearing stress is described by:

B 

PB dt

(1)

where PB is the bearing load, d is the pin diameter, and t is the thickness of the composite structure. Experimental studies correlates geometrical parameters to the resistance to failure and the fracture mode of the joints. From a theoretical point of view, in a particular geometrical condition, a given fracture mode occurs when the apparent strength of a particular mode is lower than that of other fracture modes. To determine the theoretical curves that quantitatively describe the different fracture modes it is necessary to equate the fracture loads of the three failure modes (namely: PB = PN, PN = PS, and PB = PS). In this way, it is possible to identify the geometry governing the transition among the failure modes. The bearing stress is defined in (1), net tension and shear out stress are described by:

N 

PN w  d t

(2)


S 

PS 2et

(3)

In this way three equations are obtained (Equations 4), for the transition geometry in terms of e/d and w/d, which divide the failure space into three zones characterized by the corresponding fracture modes.

w B  1 d N  e w  2 S 1 d N d

e 1 B  d 2 S

(transition : bearing - net tension) (transition : shear out - net tension) (transion : bearing - shear out)

(4‟) (4‟‟) (4‟‟‟)

To draw such curves, it is necessary to know the values of B, S, and N, which were obtained from experimental tests. The model determines three intersecting straight lines in the e/d–w/d plane, obtaining a failure map like the one shown in Figure 16.

Figure 16. Failure maps. This failure map clearly shows the influence of geometrical parameters on the failure modes. Particularly for high e/d and w/d values, bearing fracture occurs. On decreasing only the w/d ratio, a transition between bearing and net tension is observed, while increasing d with constant w, the tensile resistance is strongly reduced. Starting from the net tension failure area and decreasing the e/d ratio, a transition between net tension and shear out occurs (while a transition from net tension and cleavage is observed in experimental tests). The cleavage failure can be considered a mixed fracture. Even in this case, on decreasing e with constant d, shear resistance is reduced. Typical stress/displacement trends obtained from bearing tests are reported in Figure 17 where the value of stress, obtained by Equation (1), is recorded at increasing pin displacement.


207

In all the tests, an initial linear trend was observed with similar slopes. Different failure modes were observed to occur as a function of geometric parameters. In particular, sometimes net tension and shear out fracture modes occurred, the first at higher loads than the second with a catastrophic loss of the stress carrying capability. When a bearing failure is observed, a progressive fracture occurs which was evidenced by the presence of a plateau after the first drop of the maximum load.

Figure 17. Stress/displacement curve.

Figure 18. 3D plot of bearing stress vs w/d and e/d ratios. Figure 18 reports a 3D plot that evidences how of the stress at varying the geometrical parameters w/d and e/d. It is possible to observe:


the transition among the different failure modes identified by the failure map; - the by-dimensional plateau that characterises the bearing area. It is possible to predict the failure mode as a function both of the lamination sequence and the geometry, applying some failure criteria, specific for the composite materials, to the results obtained through Finite Element Analysis [18]. Figure 19 the stresses along the sample axis (i.e. load application load) are mapped at varying geometry parameters. In the sample of Figure 19a, bearing failure occurs. In this image a wide compressed zone was observed, above the pin. On the sides of the hole, extremely localized tensile stresses were observed, whereas the shear stresses are minor. Such observations are coherent with the fracture mode that has been observed experimentally. Compression prevails over tensile failure due to the presence of a compressed area of the loaded section which is wider than that in tension. In the sample of Figure 19b, net tension failure occurs. The compressed zone is narrower than the previous case, and it involves a region below the hole as well. Tensile stresses act on the loaded section on both sides of the hole. It is possible to observe that this section appears reduced due to the tensile load. In the sample of Figure 19c, shear out failure occurs. On both sides of the hole high shear stresses are found (coupled to low compressive and tensile ones) with a maximum value that allows the fracture to start (shear out). Such analyses confirm the validity of the theoretical model showing that changes in the geometry of the joint lead to a different distribution of stresses in the hole area; the model also shows how one of the stress components (compressive for the bearing failure, tensile for the net tension and shear for the shear out one) is preponderant on the others.

a)


209

b)

c) Figure 19. Finite Element Analysis: a) bearing; b) net tension; c) shear out. This analysis allows to foresee the stress distribution at the pin/hole contact area using as input data the physical and mechanical properties of the composite constituents, but, primarily, it allows to define a fracture predictive model by applying to the simulation results both mathematical models and failure criteria. The method [19] is based on the “characteristic curve” determination: rc    r  Rt  Rc  Rt   cos  (5) where rc is the radium of characteristic curve, Rt and Rc are respectively the tensile and compressive characteristic distances, r is the hole radius and θ is the angle around the hole. As failure criteria it is possible to use: - Hill-Tsai


  11    22    11   22    12           1  X   Y   X  X   S  2

2

-

(6)

Tsai-Wu

2 F1   11  F2   22  F11   112  F22   22  2  F12   11   22  F66   122  1

(7) where X and Y are respectively the tensile resistance in the fibres direction and in the orthogonal one, S is the shear resistance and F are the parameters dependent to the tensile and compressive resistances. Reporting in a graph (Figure 20) the characteristic curve and the line of nodes where the stress values satisfy the failure criteria, it is possible to define from the intersection an angle value that allows to identify the failure mode. It is possible to observe the following cases: 0° < θ < 15°: bearing (Fig. 1b), 30° < θ < 60°: shear out, 75° < θ < 90°: net tension, for other values the failure is mixed.

Figure 20. 0° <  < 15° - bearing failure mode.

4.2.1.2. Bearing Compression Test Figure 21 reports the test scheme [19]. The bearing tests are carried out (according to [20]) by local compression of the specimen (40×40×20 mm) in the notch using a hardened steel pin (speed: v=1 mm/min). Tests can be performed at varying the pin diameter.


211

Figure 21. Test scheme.

4.2.1.2.1. Study and prediction of failure . A typical load/displacement diagram for a compressive bearing test is shown in Figure 22. Generally, at the beginning, the graph exhibits a small nonlinearity caused by the variation in the skin/sample contact conditions; then it gradually shows a nearly linear behaviour. In proximity of the maximum load a second nonlinearity is observed until the bearing collapse arises at the skin/pin contact area. After that, the peak load is reached followed by a plateau during the fracture propagation. The load where the plateau zone starts is chosen as bearing load [21]. Figure 23 reports the damage evolution for a sandwich structure with gelcoat, at different fixed displacement levels (1 mm; 1.5 mm; 2 mm) considering a pin of 4 mm diameter. For the displacement of 1 mm, the photo shows a fracture line on the gelcoat surface that starts from the hole edge. This fracture spreads towards the sample centre across the gelcoat. After 1.5 mm, the damage is evident but the risen hole surrounding areas are still stick on the skin. Interlaminar fractures are evident near the hole edge due to the local compression. This phenomenon produces a damage of the fibres and then it reduces the sandwich stiffness. Finally, after a pin displacement of 2 mm, the gelcoat completely separates from the skin and the bearing fracture is evident on both sides of the sample.


Figure 22. Load/displacement curve.

Figure 23. Damage evolution. A Finite Element Analysis allows to identifies the region around the hole that is sensible to the bearing load effect. For example, Figure 24 shows the stress-hole distance curves at varying the pin diameter. These trends are obtained by simulating a compressive bearing test for a sandwich structure with different pins. The intersection between the numerical trend and the experimental strength of the more resistant layer (i.e. the compressive strength for un-notched specimen) defines the critical distance for each pin diameter. The distance between holes ought to be higher than two times the critical distance to avoid interference between the damaged regions of the sandwich. The stress is high near the hole and it decreases with the distance. Its reduction is more abrupt for smaller pin diameter than for larger ones, due to the extension of the sandwich pin contact surface. The critical distance increases at increasing pin diameter.


213

Figure 24. Critical distances.

4.3. Design Parameters for Bolted Joints The following parameters affect the strength of a bolted joint: 1. Material parameters such as fibre orientation, lay-up sequence, and type of reinforcement; 2. Joint parameters such as ratios of width to bolt hole diameter (w/d), edge distance to bolt hole diameter (a/d), and thickness to bolt hole diameter (t/d); 3. Quality of hole, such as delaminated edge; 4. Clamping force. The stress concentrations around the hole are reduced using doublers (local increase in thickness around the hole), minimizing component anisotropy, and using softening strips of lower modulus such as fibreglass plies in graphite composites. In multi-bolting joints, where more than one row of bolts is used, the spacing between holes (pitch) and the hole pattern are also important in designing bolted joints.

4.4. Preparation for the Bolted Joint Bolted joints are prepared by drilling holes in mating parts at specified places. Once the holes are machined, the mating parts are brought closer together and aligned to pass the bolt through the hole. A washer is placed on the other side and then the nut is placed. Bolts are often tightened by applying torque to the bolt head or nut, which causes the bolt to stretch. The stretching results in bolt tension or preload, which is the force that holds the joint together. For various applications, a pre-specified torque is applied according to the design requirement using a torque wrench and is repeated for each assembly.


References [1] [2]

[3] [4] [5] [6]

[7]

[8]

[9]

[10] [11]

[12]

[13] [14]

[15]

[16]

[17]

[18]

[19]

http://www.rina.org. Di Bella G, Borsellino C, Pollicino E, Ruisi VF. Experimental and numerical study of composite T-joint for marine application. International Journal of Adhesion and Adhesives 2010; 30: 347-358. Petrie EM. Handbook of Adhesives and Sealants. McGraw-Hill. New York, 2000. Plexus®. Guide to the Bonding. SIKA®. Sikaflex® Marine handbook: Bonding and Sealing Systems. Baldan. Adhesively-bonded joints and repairs in metallic alloys, polymers and composite materials: Adhesives, adhesion theories and surface pretreatment. Journal of Materials Science 2004; 39: 1-49. Da Silva LFM, Ferreira NMAJ, Richter-Trummer V, Marques EAS. Effect of grooves on the strength of adhesively bonded joints. International Journal of Adhesion and Adhesives 2010; 30: 735-743. Arenas Reina JM, Narbòn Prìeto JJ, Alìa Garcìa C. Influence of the Surface Finish on the Shear Strength of Structural Adhesive Joints and Application Criteria in Manufacturing Processes. The Journal of adhesion 2009; 85: 324-340. Borsellino C, Calabrese L, Di Bella G, Valenza A. Comparisons of processing and strength properties of two adhesive systems for composite joints. International Journal of Adhesion and Adhesives 2007; 27: 446-457. Mallick PK. Fiber-reinforced Composites. Marcel Dekker Inc. New York, 1993. Icten BM, Sayman O. Failure Analysis of Pin-loaded Aluminum-glass-epoxy Sandwich Composite Plates. Composites Science and Technology 2003: 63: 727– 737. Ireman T. Design of Composite Structures Containing Bolt Holes and Open Holes. PhD Thesis. Department of Aeronautics, Royal Institute of Technology. Stockholm, 1999. Collings TA. On the Bearing Strength of CFRP laminates. Composites 1982; 13: 241–252. Wang HS, Hung CL, Chang FK. Bearing Failure of Bolted Composite Joints. Part I: Experimental characterization. Journal of Composites Materials 1996; 30: 1284–1313. Quinn WJ, Matthews FL. The Effect of Stacking Sequence on the Pin-bearing Strength in Glass Fibre Reinforced Plastics. Journal of Composite Materials 1977; 11: 139–145. Kim RY, Whitney JM. Effect of Temperature and Moisture on Pin Bearing. Strength of Composite Laminates, Journal of Composite Materials 1977; 10: 149– 155. Valenza A, Fiore V, Borsellino C, Calebrese L, Di Bella G. Failure Mao of Composite Laminate Mechanical Joint. Journal of Composite Materials 2007; 41: 951-964. Di Bella G. A new failure modes predictive tool in composite laminate mechanical joint. In Proc: VIII AITEM Congress - Enhancing Enhancing the the Science Science of Manufacturing of Manufacturing; Montecatini Terme (Italy), 2007. Ramakrishna S., Hamada H., Nishiwaki M. Bolted joints of pultruded sandwich composite laminates. Composite Structures 1995; 32: 227-235.

Joining of GFRP in marine applications [20]

[21]

215

Di Bella G, Calabrese L. Pin-Contact Behaviour of Composite Sandwich Structures Under Compressive Bearing Load. Applied Composite Materials. DOI 10.1007/s10443-010-9146-y. Wu PS, Sun CT. Modeling bearing failure initiation in pin-contact of composite laminates. Mechanics of Materials 1998; 29: 325-335.