Technology of Rivet: Adhesive Joints

Abstract

In general hybrid bonding allows to match together the properties of two different joining techniques, with the aim to obtain a joint with a better set of properties (that could involve strength, energy absorption at failure, cost, etc.) with respect to the simple joints. In the literature, several works dealing with the evaluation of the strength for this kind of joints and a global enhancement of the properties, in terms of static strength, fatigue strength and energy absorption are found. The first hybrid technique “discovered” was the weld bonding and therefore in the literature it is the most deeply analyzed. In terms of mechanical performance, weld-bonding is the hybrid joining technique where better result are obtained with respect to other hybrid joining techniques such as rivet bonding, clinch bonding, etc., but at the same time it is the most troublesome in terms of joint manufacturing (influence of the presence of the adhesive on the welding performance, weld compliant adhesive requirements, etc.). On the opposite, the hybrid techniques which combine adhesive with mechanical fastening (like rivet, clinching, self piercing riveting) give in general a lower enhancement of performance if compared with weld bonded joint, but certainly they are characterized by an easier and trouble less manufacturing. In this chapter, a brief introduction to the rivet bonding technology is given, dealing with why the hybrid joints are useful and which kind of industrial requirements this kind of joints are able to satisfy. Later on, the adhesive requirements are discussed with the aim to find the best adhesive for different purposes. One of the most important benefits of the rivet bonded joints is the ease of manufacture, therefore the way to produce rivet bonded joints is discussed, and a literature example is shown with the aim to emphasize the different manufacture procedure of rivet bonded joints when compared with traditional weld bonding techniques. In this case, a noticeable benefit in terms of manufacture cost is also shown with respect to traditional joining techniques. An example of the evaluation of the strength of rivet bonded joints (pop rivet and self piercing rivet) in comparison with the performances of simple riveted and simple bonded joint, is then given. It is shown how the performance of hybrid joints depend mostly on the strength of the adhesive bond, and that the rivet becomes relevant only when the adhesive performance decreases (i.e. when the service temperature is higher than the adhesive glass transition temperature for example) or in general when the adhesive fails. Finally, the failure mode is discussed for both pop rivet bonded and self piercing rivet bonded joints. The failure of a pop rivet bonded joint is simulated using appropriate damage models for the rivet and for the adhesive. In particular the parameters of the damage models are tuned by comparison with experiments performed on simple joints and they seem to be adequate for the simulation of the failure behaviour of rivet bonded joints.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Specific Requirements in Industry

Prior to define the requirement in industry of hybrid Pop Riveted (PR) and Self Piecing Riveted (SPR) joints, a quick look to the two simple riveting techniques is given. Both are joining techniques for sheets, where the junction takes place by means of an external body called rivet. The PR is a fastening method that involves the plastic deformation of the rivet only. This allows to use this kind of joint for almost all the materials: the only need is to make a hole in the parts. Moreover, a pop riveted joint can be done even in the case the access is limited to one side of the joints.

On the other hand, SPR is a joining technique similar to clinching but a rivet is placed between the plates. Here, during the rivet setting, both the rivet and the parts are plastically deformed, therefore, brittle material cannot be joined with this technique. Furthermore, for the manufacturing of the joint, the access to both the sides is needed and a relatively powerful and expensive equipment (if compared with that of PR) is required.

Despite these advantages of PR joints, the SPR joints show typically a higher strength and durability, therefore they are used where the mechanical performances are more relevant.

The possibility to combine adhesive bonding with riveting is addressed to reduce the drawbacks of each single joining technique. For example, using the rivet-bonded joint, the strength of a bonded joint (produced with an epoxy structural adhesive) can be coupled with the rapidity of manufacturing of a riveted joint. Indeed, the presence of the rivet allows to hold the parts together during the polymerization without any other tool.

Looking at the industry, certainly one of the most relevant requirements is the joint strength. Depending on the industrial field, this requirement has to be coupled with some other features: for example in aerospace applications other relevant requirements involve the stiffness, the damage tolerance and the weight of the joint. All of these features are characteristics of adhesive bonding. If the adhesive is coupled with riveting, a quicker and easier joint production will result and at the same time the joint will also have a higher damage tolerance. In aerospace, the research trend is to tailor the choice of the material for the specific application with the aim to obtain a lighter structure [1], and it is rather common to meet joints where a strong and stiff steel is coupled with a lighter aluminium. Moreover, nowadays composite materials and polymers in general are also widely used. This leads to have in the same product several different kinds of materials: steel, aluminum alloys, polymers, composites, which can be difficult to be joined together with traditional joining techniques like welding. A solution could be bolting but it brings in a significant weight increase (absolutely to be avoided in the modern lightweight structures) and its performances are not comparable to welding. The hybrid rivet bonding technique seems to be a solution since it usually satisfies all the requirements in terms of strength, cost and weight. In this industrial field, hybrid joining in general, but especially rivet bonding is used also for patch repairs [4].

Considering the automotive, bus, truck or railway coach construction, the rivet bonding technology could be a good solution for the bodywork assembly (especially if made out of aluminum alloy) as enhancement with respect to the simple riveted joints. In fact, when parts are held using structural pop rivets, the intrinsic strength of the rivets is low, therefore butt-straps are generally needed to extend the joining area (an example is shown in Fig. 1). In this case, the number of rivets can be reduced (reducing therefore the assembly costs) adding an adhesive layer between the straps and the profile, producing at the same time a stiffer and stronger joint.

Fig. 1

Example of butt-strap riveted joint for vehicle frame applications

In the nautical industry, the coupling of adhesive and rivet attracts attention especially for the possibility to obtain a strong (by the rivet) and sealed (by the adhesive, or better the sealant) joint. Here, the materials involved are brittle (typically fiberglass), therefore pop riveting in the only possible choice. The typical geometry in this kind of application is shown in Fig. 2. In the majority of cases the joint is designed to sustain the load only by the rivets, while the adhesive has only sealing purpose.

Fig. 2

Example of hull/vessel joint in naval application

If the adhesive used is not only a sealant, but it has also good strength, the joint performance can be significantly increased.

However, regardless the specific industrial field, the most important industrial requirements involve strength and costs. Apart from this two factors, other relevant factors to be considered from the industries when a joining techniques is selected concern the materials to be joined, the energy absorption at failure (especially in the automotive crash tests) and aesthetic characteristics. All of these requirements seem to be potentially satisfied by hybrid joints.

2 Adhesive Characteristics Required

It is obvious that the mechanical performance requested for the joint affects directly the adhesive choice. If the industrial needs involve the joint strength, a stiff and strong adhesive (typically an epoxy structural adhesive) is the only possible choice. On the other hand, if the joint has a sealing purpose, a flexible but also weak adhesive (typically a sealant like silicone) has to be coupled with the rivet: in this manner the joint strength is given by the rivet while the adhesive has only sealing properties. If the joints have a sealing purpose, the classic pop rivets cannot properly be used, in fact they have a hole where the mandrel is placed. In this case SPR or a specially designed PR (an example is shown in Fig. 3) has to be used.

Fig. 3

Example of rivets for airtight applications (http://www.far.bo.it)

In this manner, a complete sealed rivet bonded joint can be obtained: this can be useful in nautical application and in the building of low-pressure tanks.

Except for a particular manufacturing procedure that will be discussed in the next paragraph, it can be said that the requirements dealing with the mechanical performance of rivet bonded joints are the only relevant requirements that have to be taken into account for the adhesive selection. This is a clearly visible benefit in comparison with other kind of hybrid joints, especially the weld bonded joints ones, where the adhesive must be properly formulated in order to permit the welding process, without an excessive burn-out or damage in general.

3 Manufacture

The manufacture of a hybrid joint consists essentially of three phases, two regarding the adhesive (adhesive deposition and adhesive polymerization) and one regarding the rivet (rivet setting): by changing the order of this three phases, different manufacturing techniques can be obtained. Weld bonded joints can be produced using a weld-through or a flow-in techniques [6]. The first is the most severe for the adhesive but is the most popular industrially, because of its rapid application in comparison with the second one which does not produce any adhesive burning, but needs time for the adhesive infiltration between the welded parts. Concerning rivet bonded joints, three manufacturing techniques can be ideally defined:

• Flow-in: the rivet is set, then the adhesive is placed and cured. In this case, a low viscosity adhesive is needed, and similarly to the case of weld bonding a long time is necessary to allow the infiltration of the adhesive between the parts by capillarity. This manufacturing technique can only be useful with the aim to increase the strength of an existing riveted joint.

• Rivet-through uncured adhesive: this procedure is essentially identical with respect to that of weld bonded joints. The main difference is that the rivet setting does not bring any adhesive degradation by heat. This is certainly the most used technique because it allows to obtain good mechanical properties (provided by the adhesive) with an easy and quick manufacturing (provided by the rivet) without any equipment for holding the parts during the adhesive polymerization.

• Rivet-through cured adhesive: the last procedure consists in the application of the rivet after the adhesive polymerization. Without any doubt, this is the less useful way to manufacture hybrid joints. In fact, all the drawbacks produced by the adhesive polymerization are similar to those shown by a simple bonded joint. The only purpose which make useful this procedure is to repair or increase the strength of pre-existing bonded joints.

From the manufacturing point of view, an interesting and comprehensive work was done by Haraga et al. [9]. This work deals with the advantages brought in by the use of pop rivet bonding. In particular, the rivet bonded joints are compared from the mechanical and economical point of view with resistance spot welded joints in case of panels assembly.

In order to understand the benefit of rivet bonded joints, their strength and manufacturing processes were compared with those of simple joints, like arc welded and simple riveted. Concerning the strength, it appears that the joints produced by the combined use of adhesive and rivets have almost the same strength as arc welding. Moreover, the strength of the hybrid joints (adhesive + rivets) is rather similar to that of simple adhesively bonded joints. This seems to be because the strength of the rivet itself is much smaller than that of the adhesive, and the force is not transmitted to the rivet since it is located in the middle of the overlap. The use of rivet bonding techniques allows to noticeably simplify the manufacturing process in comparison with the traditional arc welding, since reshaping after the distortion caused by welding and puttying to improve the aesthetic properties become unnecessary. Moreover the painting process can also be omitted when pre-coated steel is used in conjunction with rivet bonding technique.

Hybrid joining leads therefore to a significant reduction of work time (until 40% in comparison to arc welding) and costs (until 30% in comparison to arc welding), and moreover a greater reduction can be obtained if pre-painted steel is used in conjunction with rivet bonding (pre-painted steel cannot be used with arc welding).

Whit the environment in mind the energy consumption in the assembly process was also analyzed. Combined joining allows a reduction of electrical power consumption up to 84% with respect to welding.

It can be therefore stated that rivet bonded joints are cheap from the manufacturing point of view, and at the same time they give mechanical performances at least equal to the performance of traditional joining technique like resistance arc and spot welding.

4 Strength and Durability

4.1 Literature Analysis

As previously discussed, the strength and the durability of rivet bonded joints depend on what kind of adhesive is used. A first thought about the strength of a rivet bonded joint could be: “in the same joint the rivet and the adhesive coexist: therefore the strength of the rivet bonded joint will be the sum of the strength of the two simple joints”. Unfortunately this sentence is wrong. If the joint designer is smart enough, a rivet bonded joint with a strength higher than those of simple joints (but however lower than their sum) can be obtained. But sometimes, if the joint is designed in the wrong way, the performances of the hybrid joint can also be lower than those of the simple joints.

A prediction of the strength of rivet bonded joints was extensively discussed by Gomez et al. [7]: in particular a mechanical model combining springs and dampers for the representation of the behaviour of a single lap hybrid joint was developed (Fig. 4). The model is divided in two branches: one representing the adhesive and the other one representing the rivet. Each branch is made by a spring (that represents the elastic behaviour) and a damper (that represents the inelastic behaviour). This model assumes that the two parts (rivet and adhesive) work simultaneously, therefore the most rigid component will carry most part of the load, and it will be likely the first to fail, while the most compliant and ductile part will carry the load until final fracture. This model well represents the behaviour of a rivet bonded joint when the two joining techniques cooperate and they do it in the right way. With the aim to understand when an hybrid joint gives better mechanical properties than a simple joint, the joint geometry has to be analyzed.

Fig. 4

Schematic representation of the model proposed by Gomez et al. [7]

The shear stresses in the adhesive layer of a single lap bonded joint are generated by the differential deformation in the adherends [20]. Taking for example the geometry shown in Fig. 5, where L represents the overlap length, the shear stresses generated by the load P (assumed to be 100 N by mm of joint width) normalized with respect to the average shear stress along the overlap length are shown in Fig. 6 for three different values of L, as a function of the normalized coordinate x/L.

Fig. 5

Example of geometry of a single lap bonded joint

Fig. 6

Normalized shear stress versus normalized coordinate, for three different overlap length (L) for a single lap bonded joint

It can be seen that the increase of the overlap length leads to a progressive concentration of the stresses at the ends, and at the same time a progressive unloading of the zone in the middle of the overlap. This means that, in the middle of the overlap, the differential deformation is really low.

In the case of rivet bonded joints, the rivet is usually placed in the middle of the overlap. Although the rivet generate a stiffness that locally is higher than that generated by the adhesive (especially in the case of SPR joints), in the case of “long” bonded joints the differential deformations in the middle of the overlap are low and therefore the rivet contributes to the load bearing only in a negligible manner. This leads to the assumption that in a structural application, especially in the case of thin and stiff adhesives and long overlap lengths, the strength of the rivet bonded joint depends on the adhesive.

On the contrary, when the overlap is short, the mechanical fastening contributes significantly to the load bearing. Sometimes, the fastened dot brings in a reduction of bonding area in the zone where the adhesive participates to the load bearing. If this reduction is not compensated by the strength of the rivet, the hybrid joints can result in a joint weaker than the simple bonded joint. This phenomenon is not so pronounced in case of pop rivet bonded joints, where the reduction of area is relatively low, but it is significant for self piercing riveted joints, as for example shown in Fig. 7 [14].

Fig. 7

Fracture surfaces of a hybrid PR and b hybrid SPR joints

It can be seen that in rivet bonded joints (Fig. 7a) the unbonded area is essentially equal to the transverse section of the rivet, while in case of self piercing riveted joints (Fig. 7b) the contact pressure generated during the forming process spread the adhesive away, resulting in a vast unbounded area. In a recent study [13] it was demonstrated that this reduction of bonding area leads to a hybrid joint strength lower than that of a simple bonded joint.

4.2 Strength of PR and SPR Adhesive Joints

Recently, an exhaustive experimental campaign addressed to the evaluation of the strength of hybrid PR and SPR bonded joints in comparison with those of simple riveted and simple bonded joints for different environmental and geometrical conditions was done [14]. In particular, the performances were evaluated for different temperature levels, from −30°C up to 90°C, with or without an ageing treatment, for different adherend thickness and for different pitches between the rivet dots. The joint geometry is shown in Fig. 8. Two plates of galvanized steel (S 275) and aluminum alloy (AA 5052) with dimension 100 mm × 60 mm were joined together with an overlap length of 15 mm. This value was chosen taking into account the strength of the substrates in order to avoid the failure of the aluminum plates away from the overlap. The pitch of the rivet was evaluated maintaining the same joint geometry and changing the number of rivets: in order to represent different value of pitch using a joint width of 60 mm, a single rivet was placed to represent a pitch of 60 mm, while two rivets were placed to represent a pitch of 30 mm, as shown in Fig. 8.

Fig. 8

Specimens with one or two rivets, respectively, representing a pitch of 60 mm (a) and 30 mm (b)

The ageing was done following a German standard (VDA 621-415) typically used in the automotive field to test the resistance to corrosion.

The factors evaluated and their levels are summarized in Table 1. The analysis was done using a Design of Experiments (DoE) methodology, and in particular a ¼ factorial plan, with 3 repetitions for each defined treatment, was done. Since a factorial analysis was adopted, the temperature factor was split in two two-level factors as indicated in Table 2. The factor Temp 1 indicates a large variation of temperature, while the factor Temp 2 indicates a small variation of temperature.

Table 1 Factors evaluated and their levels in the experimental campaign

Table 2 Definition of temperature levels

The pop rivets used here were made of aluminum alloy (AlMg 3.5) and the geometry is schematically shown in Fig. 9. Figure 9 also shows a small table indicating the length of the rivet as a function of the thickness of the plates.

Fig. 9

Dimension of the rivet (http://www.far.bo.it)

Concerning the self piercing rivets, the machine and rivet size were changed depending on the sheet thickness, as shown in Table 3. For both the plates thicknesses, the same rivet diameter (5) and the same rivet length were used (7) and only the shape was different.

Table 3 Self Piercing Rivet joining parameters as a function of plate thickness (http://www.bollhoff.it)

For the experimental campaign, the Henkel Terokal 5077 adhesive was used: this adhesive is a pasty, heat curing one-part adhesive based on an epoxy resin. One of its main feature is to guarantee a good adhesion also to oily metal surfaces. This property makes this adhesive very attractive for several industrial fields since it allows to reduce the surface preparation that sometimes could be expensive and difficult in the manufacturing process [2, 19, 21]. Another feature of this adhesive is the possibility to be combined with other joining techniques like for example resistance spot welding. Since this adhesive was developed to guarantee high peel and impact peel resistance (load condition typically met in case of torsional and crash impact), it is particularly useful for bonding in the automotive industry.

Being a paste adhesive, its deposition at room temperature is rather troublesome; therefore, in order to reduce the adhesive viscosity, prior to deposition, the adhesive and the adherends were heated up until 50°C. This also improve the adhesion performance since the reduction of the viscosity allows the adhesive to better penetrate in the surface roughness. Immediately after the deposition of the adhesive, the parts were joined in order to avoid an increasing of the viscosity of the adhesive and a consequent reduction of the adhesion performance. The adhesive was then cured in an oven at a temperature of 160°C for 30 min.

In order to understand the influence of the temperature over the mechanical behaviour of the adhesive, tensile tests on bulk adhesive specimens were performed following the ASTM standard D 638 (Type I geometry was selected). The results of the tensile tests are shown in terms of tensile stress versus strain plots in Fig. 10. It can be seen that the adhesive performance is deeply affected by the temperature. In particular, taking as reference the strength at room temperature, at −30°C the adhesive becomes much stiffer, and it can withstand a tensile stress approximately two times that at room temperature. On the opposite, at high temperature the adhesive shows a viscoelastoplastic behaviour.

Fig. 10

Stress–strain plots of the bulk adhesive tests

The maximum strength is significantly lower than that at room temperature, although the adhesive can withstand an high elongation (until 60%) under this condition. This performance in terms of tensile test can be later compared with the strength shown by the bonded joints for the different temperatures.

Figure 11 shows the load versus displacement behaviour of tests carried out at 90°C: after failure, the elastic elongation is immediately recovered, and from the knowledge of the stress level and of the elastic modulus, the elastic elongation is supposed to be approximately 10%. Specimens were measured few minutes after the test and a residual elongation of 25–27% appears. Therefore an amount of 15–25% of viscous deformation is recovered after the failure (when the material is unloaded).

Fig. 11

Load versus displacement plot of tensile test at 90°C

Information about the glass transition temperature (T _g) of the adhesive was not given by the producer, however it can be assumed from the reduction of the Young modulus to be quite higher than 60°C.

Coming back to the performance of rivet bonded, bonded and simple riveted joints, an example of the results obtained with the single lap joint is given in Fig. 12 in case of pop rivets. Comparing the simple pop riveted and simple bonded joints, it can be seen that the latter is the strongest and the stiffest. The riveted joint shows a linear behaviour until 1,500 N, and this is followed by a significant amount of plastic deformation. This plastic deformation is completely due to the deformation of the rivet without any significant plastic deformation of the substrates, as can be shown in Fig. 13, and also demonstrated by the authors in a previous work [15]. The bonded joint shows instead a linear behaviour until 8,500 N. Here, a strong variation of the slope of the curve can be seen, corresponding to the yielding of the aluminum plate. The aluminum adherend begins to accumulate plastic deformation until failure is reached in the adhesive layer.

Fig. 12

Comparison of load versus displacement curves of simple bonded, simple pop riveted and hybrid joints

Fig. 13

Fracture surface of a simple riveted joint

The hybrid joint resulting from the combination of adhesive and rivet shows a load versus displacement behaviour that is in practice the same as that shown by bonded joint until the maximum load (only a small difference can be noticed in terms of maximum load). Immediately after the maximum load, the bonded joint suddenly fails and loses completely any load carrying capability. On the contrary, the hybrid joint shows a sudden load reduction, but it continues to withstand a load which is similar to the load carried by the rivet.

The failure mechanism can be better understood looking at Fig. 14. Here the load versus displacement plot of the riveted joint is shifted on the displacement axis of a value of displacement corresponding to the maximum load of the hybrid riveted joint. This curve is compared with that of the hybrid riveted joint. It appears that the behaviour of the riveted joint is similar to that of hybrid joint after the failure of the adhesive.

Fig. 14

Comparison of load versus displacement plot of a hybrid pop riveted joint and a simple pop riveted joint shifted of the value of the displacement at the strength of the hybrid riveted joint

Therefore, before the adhesive fails, the rivet is only slightly deformed and loaded. In the hybrid joint, the initial plastic deformation is concentrated in the plates and in the adhesive at the end of the overlap, while it is in practice negligible in the middle of the overlap where the rivet is placed. The rivet takes the load only when the adhesive fails.

The same comparison shown in Fig. 12 in the case of pop riveted joints, is shown in Fig. 15 in the case of SPR joints. Although the strength of SPR joints is noticeably higher than that of PR joints, the trend in rather similar. Alike in the previous case, the strength of the adhesive leads to the yielding of the aluminum plates. Some differences can only be noticed concerning the failure of the adhesive layer. In the case of SPR bonded joints, the fracture does not involve immediately the entire bonded region, but the presence of the SPR dot (stronger and stiffer than PR dot) contributes to a partial fracture stop: this can be noticed from the sudden drop of load close to a displacement of 2 mm (initial failure of adhesive), followed by a further load increase until complete adhesive fracture. Here the SPR dot carries the entire load and the joint behaves as a simple SPR joint until complete fracture.

Fig. 15

Comparison of load versus displacement curves of simple bonded, simple self piercing riveted and hybrid joints

A better and comprehensive view of the performances of simple and hybrid joints can be seen in Figs. 16 and 17 where the results of the analysis of variance are given in terms of the influence of the analyzed factors together with the average value of the tests.

Fig. 16

Average value and influence of factors for simple bonded, pop riveted and pop rivet bonded joints

Fig. 17

Average value and influence of factors for simple bonded, self piercing riveted and self piercing rivet bonded joints

In particular, Fig. 16 deals with the result of simple bonded and pop riveted joints and their combination. It can be seen that in average, the bonded joints are much stronger than simple riveted joints, and that the hybrid joints show an average strength rather similar to that of simple bonded. Moreover, the strength of simple riveted joints does not depend on the thickness of the plates but is obviously affected by the pitch. This is because the load is sustained simply by the rivet section and the diameter of the rivet does not depend on the plates thickness (in the rivet, only the body length is a function of the plates thickness). Since the loads reached by the simple riveted joint are relatively low, no bending of the plates is met, therefore the rotation of the overlap is negligible. This leads to a pure shear failure of the rivet that does not depend on the plate thickness. Concerning the pitch, it is obvious that a joint with two rivets will sustain a load that is close to the double of the load sustained by a joint with a single rivet. Concerning the other types of joints, the strength of the simple bonded joint is positively influenced by the plate thickness, and in a more significant manner negatively affected by the temperature and the ageing. Moving to the hybrid joints, it can be seen that the general trend is rather similar to that of the simple bonded joint. Indeed, except for the plate thickness, the relevant factors for the hybrid joints are exactly the same of those of the simple bonded joints.

The result of the analysis of variance for the simple bonded joints, SPR joints and their combination is given in Fig. 17. Again the simple bonded joint is stronger than the SPR one, although this latter is in turn significantly stronger than simple PR. In average it appears that the SPR bonded joints are weaker than PR bonded joints.

Differently from the PR bonded joints, for the SPR bonded joints the pitch assumes a significant influence. This means that in the hybrid SPR bonded joints the mechanical fastening plays a significant role in carrying the load. This fact can be confirmed looking at the influence of the temperature: the strength of the hybrid joints is less affected by the temperature than that of simple bonded joints, and this can be explained by the presence of the mechanical fastening. The same conclusion can be drawn looking at the influence of the thickness: the coefficient estimate for the hybrid joints is quite higher than that of simple bonded joints, and this is because of the presence of the SPR dot (for the simple SPR joints the thickness influence in rather high).

Using the analysis of variance (ANOVA), the strength for different geometries and environmental conditions can be predicted. Starting with the geometrical factors, Fig. 18 shows a comparison between the strength of simple bonded, simple riveted (PR) and hybrid joints (PR + adhesive) tested at room temperature as a function of the pitch and the plates thickness. It appears immediately that bonded and hybrid joints are much stronger than pop riveted joints. Moreover, it can also be noticed that the surfaces of hybrid and bonded joints are relatively flat: this means that the variation of geometrical factors slightly affects the strength. However, it can be seen that the bonded joints strength is positively affected by the thickness: in fact the higher is the thickness, the lower are the peel and the shear stress concentration at the end of the overlap, while the pitch is a parameter that cannot be evaluated for simple bonded joints. The hybrid joints show a behaviour that is rather similar to that shown by simple bonded joints: therefore, a small, but significant influence of the thickness and a negligible influence of the rivet pitch is found. This is confirmed also for different geometries: in the case of structural adhesives, the strength of hybrid pop rivet bonded joints is predominately given by the adhesive.

Fig. 18

Strength prediction of simple PR (a) and hybrid PR (b) joints, in comparison with that of simple bonded joints (dotted lines)

Figure 19 refers to SPR joints. In the same manner of PR joints, the pitch is relevant because the higher is the number of rivets per unit length, the higher is the joint strength, but here the thickness becomes also relevant. For this kind of joint, the failure corresponds to the failure of the adherends and therefore the higher is the thickness the higher is the strength (further details of the failure mode will be later given in the Sect. 5). Moving to the hybrid joints strength, it can be noticed that their strength is lower than that of simple bonded joints. This can be explained by the significant reduction of the bonded area in SPR bonded joints, in comparison with PR joints. However, it can be seen that the hybrid joints strength is mainly given by the adhesive (the hybrid joints strength values are closer to the bonded joints strength values). Differently from PR bonded joints, here a significant influence of the SPR dot pitch can be found. This means that, although the strength of SPR bonded joints is mainly given by the adhesive, the rivet also contributes to carry the load. There is therefore a kind of balance: the presence of the SPR dot reduces the bonded area hence the adhesive bond performance, but at the same time the SPR dots are strong and stiff (on the contrary of PR) and therefore they contribute to the strength.

Fig. 19

Strength prediction of simple SPR (a) and hybrid SPR (b) joints, in comparison with that of simple bonded joints (dotted lines)

So far, the behavior at room temperature has been discussed, but one of the declared benefits of hybrid joints is to compensate the loss of performance of bonded joints at high temperature. It was previously discussed how the adhesive performances decrease as the temperature increases in the case of the bulk adhesive and the same behaviour can be expected for the bonded joint. Taking as reference the room temperature performances, Table 4 shows the difference in terms of bulk tensile strength and lap shear strength as a function of temperature.

Table 4 Tensile stress of bulk adhesive specimens and predicted strength of bonded joint as a function of temperature (the variation refers to the strength at room temperature)

It can be seen that at low temperature (−30°C), the tensile strength increases significantly, but at the same time the strength of bonded joints decreases. At high temperature, the loss of strength of bonded joints is noticeably lower than the loss of adhesive tensile stress. This different behaviour can be explained looking at the properties of the adhesive. In fact, at low temperature, the adhesive is stronger than at room temperature, but at the same time it is also much stiffer and brittle. Therefore, the stress concentration at the end of the overlap is more pronounced and the failure occurs at a lower level of average shear stress. At high temperature, the more ductile behaviour of the adhesive allows a better stress distribution and therefore the loss of strength in bonded joints is lower that the loss of bulk adhesive tensile strength.

Concerning hybrid joints, their strength as a function of temperature is shown in Fig. 20 for PR bonded joints and in Fig. 21 for SPR bonded joints. It seems that hybrid and simple bonded joints have similar strengths at room temperature and 60°C, while differences emerge at −30° and 90°C. At high temperature, the adhesive becomes “soft” and its load bearing capability is reduced. The presence of the rivet allows to reduce the strength reduction, and in particular it allows to ensure a minimum level of load in case of complete adhesive failure. For the materials and geometry studied (adhesive rather stiff, relatively weak rivets, large pitch between rivets), the contribution of the rivet is not so evident, however the strength goes from 8,570 N for simple bonded joints to 9,260 N for hybrid PR joints. A significant improvement of the performance can also be noticed at low temperature. In this case, the presence of the rivet reduces slightly the peel stress within the joint and, therefore, the peel stress concentration that are critical when the adhesive is brittle.

Fig. 20

Strength prediction of PR bonded (a) and simple bonded (b) joints as a function of temperature (plates thickness 2.5 mm)

Fig. 21

Strength prediction of SPR bonded (a) and simple bonded (b) joints as a function of temperature (plates thickness 2.5 mm)

In the case of SPR bonded joints (Fig. 21), the strength still shows the loss of performance alike PR bonded joints, but the reduction is quite lower since the SPR is stronger than the PR.

Figure 22 shows the influence of the ageing for simple bonded and SPR bonded joints. In both cases the ageing affects negatively the joint strength and in particular in simple bonded joints the influence of the ageing increases when the plate thickness increases. On the opposite, it can be noticed that for hybrid joints the combined effect of thickness and ageing can be neglected, since the two lines are almost parallel.

Fig. 22

Strength prediction of bonded (a) and SPR bonded (b) joints as a function of plate thickness with and without ageing

5 Types of Failures

5.1 Fractographic Examination

As widely shown in literature and practical experience, when weak materials but also composite laminates are riveted together failure is likely to occur in the substrates [3, 12]. The most typical kinds of failure are schematically shown in Fig. 23.

Fig. 23

Typical fracture modes for fiber laminated riveted joints [3]

In this case, the failure occurs in the adherends because the load is transferred locally. This leads to the failure of the joint at a level of load that can be significantly lower than the strength of the mechanical fastening and of the adherends. Using the rivet bonded joints, the adhesive allows to transfer the load uniformly along the bonded width and, therefore, the possible kinds of failure are the failure of the adhesive and the rivet, or the failure of the entire section of the adherends. This better stress distribution permits to reach the critical condition at higher loads, and therefore the materials can be used at the best of their performances.

In case the of metal adherends, as previously shown, the failure of the hybrid riveted joints initially involves the failure of the adhesive (eventually preceded by the yielding of the adherends), and later the failure of the riveted dot. Since the two failure mechanisms occur in sequence, they are independent with respect to each other. In the rivet bonded joint, the adhesive fails as in a simple bonded joint, and the riveted dot fails as in a simple riveted joint.

Looking at Fig. 24, it can be seen that the simple bonded joint fails with a mixed adhesive fracture that propagates at the interfaces.

Fig. 24

Fracture surface of a simple bonded joint (the visible wires are those used for the control of adhesive thickness)

On the other hand, the pop rivet joint fails due to the shear failure of the rivet (Fig. 25): in particular the body of the rivet initially ovalizes, then it is sheared by the plates which act as a blade.

Fig. 25

Fracture surface of a simple PR joint

A completely different failure is shown by the SPR joints (Fig. 26). Here the rivet is made of high strength steel, therefore the failure occurs due to shearing of the plates. In this way, the strength potential of this joining technique in not completely exploited, even though this kind of failure involves a noticeable volume of material leading to a high energy absorption at failure.

Fig. 26

Fracture of a SPR joint

Coming to the hybrid joints, an example of fracture surface is given in Fig. 27. Some differences can be noticed in comparison with simple bonded joints regarding the adhesive fracture. In bonded joints, the adhesive thickness was controlled and set to 0.25 mm by using wires. On the other hand, in riveted joints, the adhesive thickness was not controlled in order to avoid the introduction of an artificial thickness variation: if during the rivet set up wires or spheres were used, a complete lack of control of the adhesive thickness would result as shown schematically in Fig. 28. Therefore, in hybrid joints, the adhesive thickness is quite smaller than in simple bonded joints, resulting in a more cohesive fracture surface.

Fig. 27

Fracture of a PR bonded (a) and SPR bonded joint (b)

Fig. 28

Gap between the plates in the case of the presence of wires during the rivet bonded joint manufacturing

5.2 Simulation of Damage and Failure

The failure of pop rivet bonded joints was simulated in detail by the authors using damage models for both the adhesive and the rivet [15]. In particular, a cohesive zone model was used for the adhesive, while for the rivet (again made with aluminum alloy AlMg 3.5) the Ductile Damage (DD) model or the Gurson-Tveergard-Needleman (GTN) models were used alternatively. The main aim of the work was to identify the damage models parameters for the rivet and for the adhesive using simple bonded and riveted joints, respectively, and later on use these parameters for the simulation of rivet bonded joints. This work is not intended to give details about the damage models, but it is focused on the hybrid joint, therefore only a brief introduction to the models is given while more details are given in the [15]. The cohesive model is a micromechanical model widely used for the simulation of fracture propagation in bonded joints or composite materials. In particular, it combines the stress and the opening at the crack tip in an energetic criterion. The parameters for this model are usually tuned by comparison with experimental tests and then used for the simulation. Assuming a trapezoidal law in terms of stress-opening behaviour at the crack tip, as shown in Fig. 29, the set of parameters is represented by:

Fig. 29

Trapezoidal stress–displacement relation of the cohesive law used for the adhesive

• The fracture energy Γ, i.e. the area underlying the curve.

• The maximum stress σ_max,

• The shape factors λ₁ and λ₂, which are the values of the displacement corresponding to the vertex normalized with respect to the maximum opening.

The ductile damage model [10, 11] is widely used for the fracture simulation of ductile materials failure: using appropriate relations it links the stress with the level of elongation and the stress triaxiality. The damage is represented as the degradation of the elasto-plastic behaviour, after the threshold plastic deformation for damage initiation $\left( {\bar{\varepsilon }_{0}^{pl} } \right)$ is attained. Beyond this point, the stress in the material is computed as

$$ \sigma = \left( {1 - D} \right)\bar{\sigma } $$

(1)

where D is the scalar damage (Fig. 30).

Fig. 30

Representation according to the ductile damage (DD) model of the stress–strain behaviour of a ductile material

The damage variable D is described as a function of the plastic displacement $\bar{u}^{pl}$ (obtained from the plastic strain by means of a characteristic length of the element L) and is related to the fracture energy, that represent the integral of the stress over the plastic displacement after the damage onset. This trend is schematically shown in Fig. 31, where an exponential law was adopted in order to reduce strain localization and instability of the analysis.

Fig. 31

Yield stress σ_y (a) and damage parameters D (b) as a function of the plastic displacement $\bar{u}^{pl}$

The GTN model is also used for the simulation of the fracture behaviour of ductile metals. The original model [8] describes the material yielding surfaces (defined by the yielding stress σ ₀) of a sphere with a void inside, in terms of stress state (in terms of average stress σ _m and equivalent stress σ _eq ) and of the void volume fraction (f) using the equation.

$$ {\frac{{\sigma_{eq}^{2} }}{{\sigma_{0}^{2} }}} + 2q_{1} f\cosh \left( {\frac{3}{2}q_{2} {\frac{{\sigma_{m} }}{{\sigma_{0} }}}} \right) - 1 - \left( {q_{3} f} \right)^{2} = 0 $$

(2)

where q ₁, q ₂ and q ₃ are parameters used to better fit the experimental results. The voids volume growth ($\dot{f}_{growth}$) is obtained from the conservation of mass by means of the plastic flow rate $\dot{\varepsilon }_{kk}^{pl} ,$

$$ \dot{f}_{growth} = (1 - f)\dot{\varepsilon }_{kk}^{pl} . $$

(3)

In this manner, the model takes only into account the growth of the voids. The model was successfully improved by [5, 16–18] in order to take into account the nucleation, growth and coalescence of voids. More details of this model can be found in Chap. 5 of this book regarding the simulation of clinch bonded joints.

The experimental tests for the identification of damage parameters were carried out on adherends and adhesive different from those of the experimental phase described before, taking into account only pop riveted joints and steel sheet plates with the geometry shown in Fig. 32, and the rivet used is that shown in Fig. 9.

Fig. 32

Geometry of the riveted joint (dimensions in mm) used to tune the parameters of the metal damage models

The adhesive used for the hybrid joints for tuning models is the Hysol 9466, a two component structural epoxy resin supplied by Henkel. This adhesive has a curing cycle of 24 h at room temperature.

For the characterization of the cohesive zone model, mode I fracture tests on Double Cantilever Beam (DCB) joints were used and the set of parameters identified is shown in Table 5. As a first approximation, a mode mixity independent behavior was assumed for the adhesive (i.e. the mode II parameters were assumed to be equal to the mode I parameters).

Table 5 CZ parameters for the adhesive Hysol 9466

Concerning the simple riveted joints, finite element simulations were carried out for the tuning of the ductile damage parameters. The finite element model is shown in Fig. 33. The plates and the rivet were modeled with 3D deformable elements while analytical rigid surfaces were used to represent the rivet mandrel and the reference surface: in order to take into account the residual stresses that occur during the rivet set up, the forming of the rivet was also simulated. More details on this choice are given in [15] while here only the principal results are discussed.

Fig. 33

Finite element model of the riveted joint

Both the parameters of the DD and the GTN models were identified using an inverse method based on the fitting of the experimental load-elongation of simple riveted joints. Concerning the DD model, different combinations of parameters were tested, and the best results are shown in Fig. 34. The numerical load versus displacement curves shown in Fig. 34 are all in good agreement with the experimental ones. Figure 35 shows some stress contour maps on the deformed shape of simple PR joints. Figure 35a is before any displacement is applied, therefore it shows the residual stresses due to rivet forming. The rivet is then loaded until the onset of damage (Fig. 35b). The damage is greater where the maximum deformation occurs and, therefore, in the green-coloured zone in the middle of the rivet. Figure 35b corresponds to a value of displacement of 0.6–0.7 mm in the graph of Fig. 34. In Fig. 35c, the damage propagates to the entire section and final fracture occurs (blue-coloured elements mean completely damaged elements).

Fig. 34

Comparison between experimental and finite element analysis load–displacement curves of the riveted joint obtained with the DD model

Fig. 35

Simulation of the shear test of the simple riveted joint: a residual stress after forming of the rivet; b onset of damage; c failure

The same procedure was followed for the simulation using the GTN model. The results are shown in Fig. 36 and the set of parameters obtained is given in Table 6.

Fig. 36

Comparison between experimental and numerical load–displacement curves of a riveted joint obtained with the GTN model

Table 6 Parameters used for the GTN model (riveted joint)

An interesting comparison can be made about the fracture surface of the rivet. Figure 37 shows a comparison between the experimental fracture surface of a riveted joint and that resulting from the simulation, where the damaged elements have been removed. It is evident that the simulation is able to reproduce closely the fracture surface obtained experimentally. In both experiment and simulation, two patterns can be noticed: in the load direction (position A in Fig. 37), the surface is flat because it is sheared by the plates, while perpendicularly to the load direction (position B in Fig. 37) the surface is rougher since it undergoes a significant deformation before failure.

Fig. 37

Comparison between fracture surfaces obtaines experimentally (a) and by the FE simulation (b, c)

Failure of the hybrid joints can be therefore simulated by assigning the parameters of the damage models calibrated previously on simple joints, to the different constituents of the hybrid joints.

Figure 38 compares the load–displacement simulations of the PR bonded joint using the DD model, the GTN model and the results of the experimental test. In both simulations, the CZ model was used to simulate the adhesive layer. A good agreement is found between the two simulations and the experimental test, especially regarding initial stiffness, displacement at failure and behaviour after the failure of the adhesive layer, which occurs close to a displacement of 0.4 mm. The peak load in both simulation is only about 12% lower than the experimental one.

Fig. 38

Numerical load versus displacement curves from simulations of the hybrid PR-bonded joint using the DD model and the GTN model

The difference between the experimental and the simulated behaviour can be related to the use of the same cohesive zone parameters for mode I and mode II, while adhesives likely show a mode II fracture toughness higher than mode I fracture toughness. The load oscillations after the failure of the adhesive are due to the sudden release of strain energy when the adhesive fails. In Fig. 39, the progressive growth of the adhesive failure in the rivet bonded joint is shown. The fracture initiates at both the ends of the overlap and propagates toward the center of the joint. When the adhesive is completely failed, the load is sustained by the rivet, that behaves as a rivet in a simple riveted joint.

Fig. 39

Progressive failure of the adhesive layer for the PR bonded joint simulated using the DD model

It appears therefore that either DD or GTN damage models can be used for the simulation of the PR bonded joint and, moreover, that the deformation and failure behaviour of the hybrid joint can be properly modeled using the damage models identified previously on simple riveted and adhesive joints.

6 Examples of Use

The hybrid joining technique is used when some requirements cannot be obtained by a single joining technology: this could involve the strength but also some other factor as cost or the feasibility in general. Therefore, a rivet bonded joint could be used as replacement or repair of simple bonded joints with an extremely simple manufacturing procedure, although it does not give the same performance of a welded, or a weld bonded joint. Rivet bonding is a good choice when different materials have to be joined (aluminum, magnesium, steel and composite material), or in case of aesthetics requirements. Some other relevant applications concern aircraft and automotive panel fabrication and the assemble of truck cabs, and in general in other automotive aluminum frames due to its capacity to withstand different kind of dynamic and impact loads. Rivet bonding is also used for the internal structures of aircrafts. Several examples of application of the rivet bonding technology can be found: two of them concern the automotive and the bus construction industry. Rivet bonding is indeed largely used for the construction of the aluminum monocoque of the XJ Jaguar series car: here the use of hybrid techniques led to a weight reduction in the order of 40% and to a stiffness enhancement of 60% with respect to the previous models. The rivet bonding is also used for the assembly of school buses (Fig. 40), as replacement of traditional simple riveting. In particular, a strong reduction of the number of rivets can be obtained producing a noticeable reduction of assembly cost. Moreover, the adhesive also works as a sealant, leading to a significant reduction of the presence of leaks, to a better vibration damping and to an aesthetic improvement.

Fig. 40

US school bus: an example of application of the rivet bonding technology

Moving to shipbuilding industry the rivet bonding is widely used for sealing purposes, especially when the strength of the sealant is not sufficient. In this case the rivet is used to sustain the load and silicone or polyurethane are the best choice for the sealing purpose.

7 Conclusions

In this chapter the technological features of rivet bonding technology are discussed: initially their advantages are examined in comparison with the industrial requirements with the aim to define the most worthwhile fields of application. Later on, the manufacturing processes are shown, with particular emphasis on the comparison with other traditional joining techniques like welding. The strength of hybrid rivet-bonded joints in comparison with that of simple bonded joints is discussed for different geometrical and environmental conditions. For both the pop rivet (PR) bonded and self-piercing rivet (SPR) bonded joints, it appears that at room temperature the performance is predominately given by the adhesive (a strong and stiff epoxy adhesive in this case), while the mechanical fastening significantly takes part to carry the load when the joint is subjected to high temperatures. Since the SPR is quite strong, its contribution to the hybrid joint strength is correspondingly high, even though the SPR set up process leads to a significant bond area reduction. In order to better understand the mechanical behaviour of hybrid joints, their failure mechanism are discussed, and in particular the failure of pop rivet bonded joints is simulated using appropriate damage models for the adhesive and the metal parts. Finally, some example of use of rivet bonded joints are shown, giving emphasis to the expected advantages.

The rivet bonding technology appears as a valid alternative to the traditional joining techniques, especially when different materials such as aluminum alloy and steel have to be joined giving the strength of the adhesive bonding but with an easier manufacturing process.