Journal of Manufacturing Processes 35 (2018) 254–260 Contents lists available at ScienceDirect Journal of Manufacturing Processes journal homepage: www.elsevier.com/locate/manpro Durability of orbital riveted steel/aluminium joints in salt spray environment T ⁎ G. Di Bellaa, , C. Borsellinob, L. Calabreseb, E. Proverbiob a b NAVTEC, c/o CNR ITAE, Via Comunale S. Lucia 40, Messina, 98125, Italy University of Messina, Contrada di Dio, Messina, 98166, Italy A R T I C LE I N FO A B S T R A C T Keywords: Joining Durability Orbital forming Corrosion In this work, aluminium alloy/steel joints were realised with an orbital riveting process. Two series of joints were made; i.e. in the ﬁrst, the head of rivet touches with the aluminium sheet; in the second, it touches with the steel one. After their manufacturing, during tests were carried out by exposing the joints to a salt spray fog up to 10 weeks into a climatic chamber. Finally, mechanical tests were performed with the aim to know the eﬀect of corrosion on performance and failure. 1. Introduction Nowadays, industries develop innovative leading technologies to respond to market requirements . In joining technology, riveting is signiﬁcant in various industries (i.e. especially in aeronautical industry and, particularly, in the connection among the sheets in aircraft). Therefore, many progresses have been done for its improvement in terms of performances (i.e. resistance and durability of the joint). The process of riveting presents diﬀerent procedures; i.e. pop riveting, press riveting, explosion and radial frictional riveting . The main critical issue is the resistance and the sensitivity to failure. In particular, this depends directly on tension status of the joint. Various parameters (i.e. riveting load, friction ratio and tolerance) affect directly on the stress and the fatigue life of the riveted joint. To increase the joint properties, in pressing operations, Collette et al applied non-uniform-impact force . In such methods, study on rivet microstructure , evidenced the starting of numerous cracks and the dissolution of rivets’ structure. Moreover, after applying desired forces, it much needed to consider tension , torsion tolerance  and residual stress in work pieces. According to experiments conducted on sheets connected by changing the type of riveting, sheets riveted by pressing operations  respect to modern methods, are characterised by a less dissolution of rivets’ particle structure. A new vision in manufacturing, including just-in-time (JIT) manufacturing and measurable process control and the requirement of a joint with less residual stress, increased the use of a new technology: the orbital forming. This is a cold forming process, alternative to conventional fastening ⁎ operations (i.e. staking, peening, crimping, pressing, swaging, spinning, rolling, riveting, welding, upsetting) . It is comparable to impact and compression forming, where is applied with a tool a compressive axial load to deform the piece. The diﬀerence with the previous processes is that, in orbital forming, the tool rotates at a ﬁxed angle (i.e. typically 3° to 6°) and applies axial and radial forces to progressively deform material until the speciﬁc shape is reached (see Fig. 1 ). Moreover, the process requires more tool revolutions and typically takes 1.5–3.0 s to complete. During the process, the deformation work interests only the tool/ rivet line of contact, not the whole tool surface. This fact reduces axial loads of about 80% by inducing several advantages; i.e.: • lower level of stress on the parts that have to be fastened or mated; • smooth surface ﬁnish; • elimination of cracks caused by impact riveting; • cold-head forming by avoiding bending or swelling of the fastener shank; • use of smaller presses in terms of sizes and costs; • less rigid ﬁxturing and longer lasting tools. The process is employed with diﬀerent materials; i.e. metals (ferrous and nonferrous) and plastics . However, the industrial applicability of this joining technology is strongly limited by highly aggressive environmental conditions that can induce localised corrosion mechanisms : i.e. the junction between dissimilar substrates (i.e. steel/aluminium) can induce corrosion Corresponding author. E-mail address: email@example.com (G. Di Bella). https://doi.org/10.1016/j.jmapro.2018.08.009 Received 17 April 2018; Received in revised form 11 July 2018; Accepted 8 August 2018 1526-6125/ © 2018 Published by Elsevier Ltd on behalf of The Society of Manufacturing Engineers. Journal of Manufacturing Processes 35 (2018) 254–260 G. Di Bella et al. Fig. 2. Geometry of substrate [mm]. Fig. 3. Geometry of rivet [mm]. The rivet was realised with a S11aluminium alloy. Fig. 1. Orbital forming. 2.2. Geometry phenomena for galvanic eﬀects. In fact, these metals have a quite different electrochemical behaviour . This phenomenon can be intensiﬁed by crevices (i.e. crevice corrosion attack) or irregularities. For this fact, in joining design, the metal sheets have to be chosen by evaluating also electrochemical and corrosion properties . Moreover, internal stresses can inﬂuence the durability. This can induce starting and diﬀusion of local cracks (i.e. eﬀect of stress corrosion cracking - SCC), by reducing the joint resistance by increasing the ageing times . Porcaro et al.  shown that the mechanical strength of aluminium riveted joints is stable after 3 days of natural ageing. Moroni et al.  investigated the behaviour of hybrid adhesive-mechanical joints after thermal cyclic ageing. Calabrese el al.  showed that, after long durability time (i.e. 60 days), the mechanical performance of aluminium self-riveting joints decreases signiﬁcantly, by evidencing that the corrosion phenomena inﬂuence performance and failure. Although the durability of dissimilar joints in a corrosion fog is well known, few works operate on the ratio between joint durability and electrochemical characteristics of the metal [13,14]. In particular, the goal of this research is to study the performance of a hybrid joint between an aluminium sheet and a steel one, realised with an orbital forming process by focusing the attention not only on the mechanical resistance but mainly on the durability in salt spray fog. This work follows other studies performed by the Authors that in recent years have investigated several joining techniques between dissimilar materials: i.e. self-piercing riveting [17,14], clinching and , clinch-bonding . In Figs. 2 and 3 are reported, respectively, the geometry of the substrate and the rivet. The thickness of the substrate is 1 mm for steel alloy and 2 mm for aluminium one. Two conﬁgurations of joints have been made: in the ﬁrst, the rivet part subjected to orbital forming touches with the aluminium sheet (in the next, series A); in the second, it touches with the steel (in the next, series F). 2.3. Joining process Orbital forming was performed using a BK-TAUMEL “BK80” machine (Fig. 4). Its characteristics are reported in Table 1. Table 2 reports the setup parameters; i.e. F is the punch force that deforms the rivet, t is the working time and Δx is the displacement of the punch. 2. Experimental setup 2.1. Materials The substrates of the joint were realised using, respectively, a 6082 aluminium alloy, subjected to a heat treatment process (i.e. T6), and a steel alloy A570. 6082 aluminium alloy is characterised by a good strength and a really good corrosion strength. It is the better of the 6000 series alloys and, for this fact, is used mainly in structural applications. In plate form, the alloy is used for machining. Carbon steel A570 is widely used in production. It has good corrosion strength, high harness, toughness and strength. Fig. 4. BK-TAUMEL “BK80” machine. 255 Journal of Manufacturing Processes 35 (2018) 254–260 G. Di Bella et al. Table 1 Characteristics of the BK-TAUMEL “BK80” machine. Technical features BK80 Max rivet shaft diameter (with material strength 400 N/mm2) Stroke min - max Max push strength Motor output Air pressure Air consumption 8 mm 0–40 mm 13.5 kN 0.37 KW 5.2 bar 10.4 Nlt Table 2 Setup parameters. F [kN] t [s] Δx [mm] 8 4 0.85 Fig. 5. Typical load/displacement curves for un-aged samples. 2.4. Corrosion treatment After the orbital riveting, the samples were placed into a climatic chamber and corrosion tests are performed in according with ASTM B 117 standard. In particular, the main conditions are: i) a 5% NaCl solution with a pH of 6.5–7.2 as salt spray fog and ii) a temperature of 35 °C. This creates a controlled corrosion environment that was used to obtain important information about the corrosion behaviour of the exposed joints. After each week, 5 specimens for each series were removed and tested. Such samples were cleaned out, dried and stored in plastic bag with desiccant silica gel. 70 samples has been realized (ﬁve for each series and for each time interval). The aim of corrosion testing is to evaluate how the fracture mechanisms of hybrid joints are aﬀected by the environmental conditions where they work. Therefore, the corrosion mechanisms within the joint were investigated to study the ratio between the corrosion degradation mechanisms and the failure ones. • A similar discrimination in three main stage can be deﬁned for the load displacement trend observed in series F. In particular: • Concerning the ﬁrst stage, after the mechanical set-up adjustments, 2.5. Single lap shear test • Mechanical tests were performed in according to ISO/CD 12996, by an UTM (Zwick-Roell Z600) with a 50 kN load cell. The cross-head rate is equal to 1 mm/min. 2.6. Analisys of joints • To analyse the corrosion eﬀect at the materials interfaces, the joints were cut obtaining the Cross-section proﬁles. The cross section was investigated by a Zeiss Stemi 2000-C stereomicroscope. 3. Results and discussion In Fig. 5, typical load/displacement curves related to the un-aged samples for both series are reported. For the joints of series A, it is possible to identify several regions: • Initial region. In this stage two sub-steps can be identiﬁed. Firstly, • contact for eﬀect of a bending deﬂection at the ends of the sheets caused by the joint asymmetry. In this phase, the performance of the joint depends only on the rivet and the resistance is bear by the joining point. Secondly, the load reaches the maximum value, and by increasing the deformation of the joint, a plateau is shown evidencing a progressive joint failure. This region is signiﬁcant because of a bearing phenomenon around the button on steel side produced by a sliding action, where the steel sheet progressively and plastically deforms, the steel hole loses its circularity (Fig. 6). Residual resistance stage. The ΔP/ΔL ratio reduces over 50% respect its maximum value by evidencing that the joint results critically damaged. The residual strength progressively decreases, a drastic reduction of the load is not evident, but it is possible to notice a gradual evolution of the damage. Finally, for higher deformation, a complete joint fracture occurs for net tension with a drastic load reduction (Fig. 6). we observe a linear relationship between load and displacement, related to joint stiﬀness. The curve slope is quite similar to that noticed for the series A. Maximum load stage. In this stage, the non-linear trend is not evident and, after the stage I, the load reaches maximum value that is lower than the one observed in the series A. This phenomenon takes place for the start of cracks induced by the orbital forming on the steel sheet. The higher workability of aluminium avoids the creation of these cracks on the samples of the series A. Residual resistance stage. The ΔP/ΔL ratio reduces over 50% compared with its maximum value by evidencing that the joint results critically damaged. Load progressively decreases for the diﬀusion of the cracks but the failure does not drastically occurs due to the bearing around the button on steel side - caused by a sliding action – (Fig. 6). This failure mechanism is in competition with the cracks’ propagation. The load slightly increases and, then, it decreases again with a gradual evolution of the damage. This causes the complete joint fracture that takes place exclusively for bearing. Further information concerning the mechanical properties of hybrid joint conﬁguration can be obtained by analysing in detail Fig. 6. In particular, although both the series shown a fracture that interests the steel sheet, the mechanical collapse is aﬀected by the joint conﬁguration. In the joint A, where the steel sheet is between the rivet head and the aluminium one, the fracture occurs for mixed bearing and net tension mode. This failure mechanism could be caused by the interlocking that avoids the unbuttoning of steel sheet. Thus, a progressive bearing plastic deformation of steel sheet is evident in the hole area until the collapse for net tension. Whereas, in the joint F, the steel sheet the trend is related to mechanical adjustments (i.e. presence of clearance and oﬀ-axis). This sub-step is not representative and it is not considered for performance evaluation of orbital riveted joints. In the second sub-step a load linear increase versus displacement is observed; i.e. the joint strength is due to the shear resistance of the rivet. Maximum load stage. Also in this case, two sub-steps can be deﬁned. Firstly, the load–displacement curve become not-linear with a progressive reduction of the slope ΔP/ΔL. The sheets progressively loses 256 Journal of Manufacturing Processes 35 (2018) 254–260 G. Di Bella et al. Fig. 6. Typical failure mechanisms of un-aged samples. unbuttoning is possible by considering the geometry conﬁguration. In fact, a local plastic deformation in the hole area, due to an extensive bearing collapse of the steel sheet, is evident. The unsymmetrical conﬁguration of the single lap joint conﬁguration induces a joint twisting that potentially, at longer displacement, can induce an unbuttoning along the rivet shaft. These considerations can be conﬁrmed by considering that the joints of the two series are characterised also by different values both of maximum load and displacement. In particular, the curve of series A is characterised by a wider area under the curve. Then, these joints requires a higher energy before reaching the failure. This behaviour is a consequence of the diﬀerent failures mechanisms, observed in Fig. 6. In Fig. 7 are reported the typical load/displacement curves related to the samples by varying the ageing time for both the series. In particular: • For series A (Fig. 7a): the curves present diﬀerent trends by chan• ging the time of corrosion. It is possible to evidence three groups: (i) the curve of the un-aged sample; (ii) the curves of the treated samples after 1, 3, 4 and 6 weeks; (iii) the curves of the treated samples after 8 and 10 weeks. For series F (Fig. 7b): the curves present similar trends where the Fig. 8. Analysis of failure mechanisms: (a) series A; (b) series F. one related to the un-aged sample gradually modify in the curve related to the treated sample after 10 weeks This can be explained by analysing the failure mechanisms by changing the time of corrosion, as evidenced in Fig. 8 for both the series. As for the series A, when the sample is not corroded (i.e. week 0), the failure occurs for bearing and net tension (Fig. 6). The cracks develop from the hole along an orthogonal direction than the load axis. Moreover, it is evident also the bearing along the load axis on the steel sheet. At this stage the steel become the critical substrate of the joint. Between 1 and 6 weeks, the failure takes place mainly for unbuttoning and, in few cases, for bearing (Fig. 9). The unbuttoning is characterised by three steps: (i) the sample bends because of the deformation of the sheet; (ii) this last causes the load transfer to the rivet head until its failure; (iii) the sheets pull out of the rivet shank. The bearing interests the steel sheet. After 8 weeks the failure mechanism is diﬀerent. Failure occurs mainly for shear out of the aluminium sheet (Fig. 9). Occasionally, bearing was also observed. This is due to the degradation mechanism occurring during ageing time. Aluminium sheet acts as anode and steel sheet as cathode for eﬀect of the galvanic coupling. Thus, a rapid Fig. 7. Typical load-displacement curves by varying the corrosion time: (a) series A; (b) series F. 257 Journal of Manufacturing Processes 35 (2018) 254–260 G. Di Bella et al. Fig. 9. Typical failure mechanisms – Series A. Fig. 10. Typical failure mechanisms – Series F. maximum load with its standard deviation is reported by increasing ageing time. In such Figure it is evident that the mean maximum load is higher, after 1, 3 and 4 weeks, than the untreated sample. This fact is caused by the formation of corrosion products that in an initial phase promote the joining (i.e. interlocking eﬀect [18,19]). Then, by increasing the corrosion time, the maximum load decreases due to the damage of substrates and rivets. In addition is it evident that the joint of series F present lower mechanical properties, due to the diﬀerent deformation of the sheet steel. To know how the corrosion aﬀects the samples’ behaviour, the section of the joints were investigated. In particular, Fig. 12 reports the sections of all the joints by varying ageing time and series. In Figure are evident the corrosion products. Their volume increases by increasing the ageing time. The steel corrosion products are brown and red, whereas the aluminium corrosion products are white. In the steel/aluminium contact region is evident the galvanic eﬀect. In fact, this induces the dissolution of aluminium and its thinning, as evident at the week 10. The steel shows only the creation of a slight layer of oxide. Speciﬁcally, for series A: dissolution of anodic metal near the contact region takes place by inducing a premature failure in the hybrid joint. By analysing joints’ failure mechanisms of series F, the diﬀerent conﬁguration (i.e. the rivet part subjected to orbital forming touches with the steel sheet) induces another trend in the failure modes. In this case, bearing and unbuttoning are in competition (Fig. 10). For untreated samples and for low corrosion times, failure occurs mainly for bearing. Whereas, for high corrosion times, failure occurs mainly for unbuttoning by following the same scheme described for the samples of series A. Despite series F, where a progressive evolution from bearing to unbuttoning of the steel sheet occurs, the series A shows a very high sensitivity of the failure modes to the ageing time. In fact, ﬁrstly, at low ageing time, a transition from bearing/net tension of steel sheet to unbuttoning of aluminium sheet takes place. At long ageing time, the shear out of aluminium sheet became the dominant failure mechanism. This behaviour is strictly related to the corrosion phenomena. At low ageing time the creation of corrosion products at the steel/aluminium interface progressively reduces the joint interlocking forces by stimulating the joint unbuttoning. At long ageing time the galvanic coupling is the cause of the aluminium degradation and, as a consequence, the aluminium thin and it breaks for shear-out. To better evidence the diﬀerence, both between the two series and within each one, regarding the mechanical joint resistance, in Fig. 11 • After 1 week: It is evident the creation of the corrosion products • Fig. 11. Maximum load by varying corrosion time. 258 between the sheets where there is maximum contact between the diﬀerent materials (i.e. galvanic corrosion). Moreover, the interstice is really low by promoting a corrosion for crevice. The corrosion products present high volume and the sheets deform. It is possible to notice also the pitting localised corrosion caused by the damage of the oxide layer on aluminium (box A). The chlorides into the salt spray fog damage the weakest points of this layer. The corrosion products ﬁll up the gap between sheets and rivet by promoting the interlocking phenomenon (box B). After 4 weeks: It is evident the high increase of volume of the products. The corrosion phenomenon interests mainly the aluminium (box C). The galvanic corrosion, coupled with pitting and crevice, induces the disappearance of an entire surface layer of aluminium. It is possible to notice also the creation of corrosion products between the formed rivet and the aluminium due to the diﬀerent alloys that constitute the two parts (box D). Journal of Manufacturing Processes 35 (2018) 254–260 G. Di Bella et al. Fig. 12. Sections of joints. • After 6 weeks: Corrosion products further increase and the rivet • • failure occurs for bearing (low corrosion time) and unbuttoning (high corrosion time). For both series, at low corrosion time, the strength of the joints slightly increases for eﬀect of the interlocking phenomenon. The corrosion test evidenced that joints of series A are characterised by better performance because the rivet part subjected to orbital forming is highly preserved. head presents localised corrosion for pitting caused by the chlorides that attack the oxide layer (box E). Moreover, the aluminium tins due do the metal dissolution caused by the galvanic coupling. After 8 weeks: In the photo is evident the thinning of the aluminium steel and the anodic behaviour of this last preserves the alloy of the rivet to resists corrosion. After 10 weeks: It is possible to notice the galvanic corrosion of the rivet head that touches with steel substrate. References Similar analysis for the joints of the series F leads to ﬁnd that the corrosion of the rivet part subjected to orbital forming is more evident.  Gunnarsdóttir SA, Rodríguez Basurto A, Wärmefjord K, Söderberg R, Lindkvist L, Albinsson O, et al. Towards simulation of geometrical eﬀects of laser tempering of boron steel before self-pierce riveting. Proc: CIRP 2016;44:304–9.  Mahmoudi M, Sadat Baha M, Khajehpour P, Youseﬁ A. Experimental and analytical comparison in orbital riveting machine and radial riveting unit in an orbital machine. Proc: 3rd International Conference on Research in Science and Technology. 2016.  Collette Q, Sire S, Wouters I. Lap shear tests on repaired wrought-iron riveted connections. Eng Struct 2015;85:170–81.  Huang L, Shi Y, Guo H, Su X. Fatigue behavior and life prediction of self-piercing riveted joint. Int J Fatigue 2016;88:96–110.  Choi JH, Yoo SY, Kim CH, Kweon JH. The structural analysis and strength evaluation of the rivet nut joint for composite repair. Compos Struct 2016;136:662–8.  Skorupa M, Machniewicz T, Skorupa A, Korbel A. Fatigue strength reduction factors at rivet holes for aircraft fuselage lap joints. Int J Fatigue 2015;80:417–25.  Casalino G, Rotondo A, Ludovico A. On the numerical modelling of the multiphysics self-piercing riveting process based on the ﬁnite element technique. Adv Eng Softw 2008;39:787–95.  http://www.taumel.com.  http://nebinfo.com. 4. Conclusions From the analysis of the data from experimental tests conducted on hybrid steel/aluminium joints it is possible to observe that: the mechanical behaviour of the joints, realised by orbital forming, are inﬂuenced by both the corrosion time and the conﬁguration (i.e. position of the rivet). The joint conﬁguration aﬀect both the failure modes for the untreated samples (bearing/net tension of A versus bearing/potentially unbuttoning of F) and the total resistance (i.e. joints of series A are characterised by higher failure loads). By increasing corrosion time for series A failure takes place for unbuttoning (1–6 w) and shear out (8–10 w). In the joints of series F the 259 Journal of Manufacturing Processes 35 (2018) 254–260 G. Di Bella et al.  Jagtap S, Kavade M. Orbital Riveting – a design and development of new machine. Int J Mech Prod Eng Res Dev 2014;2:6–10.  He X, Pearson I, Young K. Self-pierce riveting for sheet materials: state of the art. J Mater Process Technol 2008;199:27–36.  Bardal E. Corrosion and protection. Berlin: SprigerVerlag; 2004.  Krüger L, Mandel M. Electrochemical behaviour of aluminium/steel rivet joints. Corros Sci 2011;53:624–9.  Calabrese L, Bonaccorsi L, Proverbio E, Di Bella G, Borsellino C. Durability on alternate immersion test of self-piercing riveting aluminium joint. Mater Des 2013;46:849–56.  Hoang NH, Langseth M, Porcaro R, Hanssen AG. The eﬀect of the riveting process     260 and aging on the mechanical behaviour of an aluminium self-pircing riveted connection. Eur J Mech – A/solids 2011;30:619–30. Moroni F, Pirondi A, Kleiner F. Experimental analysis and comparison of the strength of simple and hybrid structural joints. Int J Adhes Adhes 2010;30:367–79. Calabrese L, Proverbio E, Di Bella G, Galtieri G, Borsellino C. Failure behaviour of SPR joints after salt spray test. Eng Struct 2015;82:33–43. Calabrese L, Proverbio E, Galtieri G, Borsellino C. Eﬀect of corrosion degradation on failure mechanisms of aluminium/steel clinched joints. Mater Des 2015;87:473–81. Calabrese L, Galtieri G, Borsellino C, Di Bella G, Proverbio E. Durability of hybrid clinch-bonded steel/aluminum joints in salt spray environment. Int J Adv Manuf Technol 2016;87:3137–47.