Impact of Rivet Head Height on the Tensile and Fatigue Properties of Lap Shear Self-Pierced Riveted CFRP to Aluminum.
Vehicle mass reduction is an important strategy to improve fuel efficiency and reduce emissions in passenger cars and trucks. Fiber reinforced polymers are becoming increasingly attractive alternatives to traditional structural automotive materials due to their potential to achieve a high strength-to-density ratio . In particular, structural and non-structural components made of carbon fiber reinforced polymer (CFRP) composites are considered for use in passenger vehicles. The introduction of CFRP in an automotive body structure can result in weight savings up to 10% greater than the use of aluminum and other lightweight metals, and can contribute up to 50% weight savings when compared to a steel structure [1,2].
One major challenge of CFRP implementation in automotive structures is joining. CFRP structures may be joined to other composite components, lightweight metals, or steels. Traditionally, resistance spot welds (RSW) have been used extensively in the automotive industry, but cannot be applied to thermoset polymer materials. Other alternatives for joining composites include mechanical fasteners and adhesives which have cycle-times that are not conducive to mass production. Recently, self-piercing rivets (SPR) have emerged as an economical and effective technique to join lightweight metals, and they have been widely used by auto manufacturers to join aluminum sheets for vehicle bodies [3, 4, 5] SPR joints between metal sheets have shown superior peel and fatigue strength compared to RSW joints [3,6, 7, 8, 9, 10]. SPRs require neither heating of materials, nor pre-processing steps, which make them ideal for mass-market automotive assembly cycle times. Recent studies have indicated that a SPR can be effectively used to join fiber reinforced plastics to metals [11, 12, 13, 14, 15].
SPR joints are best characterized by the deformation and subsequent interlock between the rivet and substrates. In metal to metal SPR joints, the amount of rivet engagement, which can be characterized in part by the rivet head height, has been shown to play a critical role in the quasi-static strength and fatigue performance of the joint [16,17]. While prior studies on SPR joints between metals and fiber reinforced composites focused on the joint quality affected by die geometry , in-plane distance between the rivets , and oil pressure in a hydraulic system [20,21], no literature is available on the strength and fatigue performance of SPR joints in composites and metals with consideration of the rivet head height (and corresponding extent of rivet engagement). However, this effect is important to quantify as a head height range must be identified that can produce joints with adequate performance.
Hence, in the present study, the effort is focused on determining the effect of rivet head height on the mechanical properties of a lap shear joint configuration, including the quasi-static failure load and fatigue performance of SPR joints in dissimilar CFRP to AA6111. Two distinct rivet head heights chosen for the study are based on a previous proprietary investigation to determine insertion depths, as well as rivet and die combinations, which would result in adequate rivet engagement for the considered stacks. Joints produced with these two rivet head heights are then compared to SPR joints produced in two similar sheets of AA6111. Finally, by utilizing the fatigue test results from this study, the analytical structural stress model developed by Rupp et al.  was used to evaluate the performance and develop a master curve.
MATERIALS AND EXPERIMENTAL PROCEDURES
Tested stacks were composed of a 2.5 mm continuous braided fiber CFRP with a thermoset epoxy resin matrix (top sheet) to 2.5 mm AA6111 as shown in Figure 1a, and a homogenous two sheet stack of 2.5 mm AA6111. In all cases, the AA6111 was in the T4 condition when the joints were created. Composite coupons were cut from compression molded plaques using a wet diamond saw. The axial fiber orientation was always transverse to the longest coupon dimension. Aluminum coupons were sheared from blanks with the rolling direction parallel to the longest coupon dimension.
A servo-driven Henrob SPR gun was used to set Henrob tubular-style rivets shown in Figure 1b. The same rivet and die combination was used for all specimens tested in the present work. The rivet insertion process was controlled via adjustment of the punch velocity. The velocity was tuned in 5 mm/s increments to deliver the desired rivet head height or flushness. Specimens were produced with a flush rivet (no head protrusion) in the mixed material and similar material stacks, and with a head height 0.30 mm above the surface in the mixed material stack. The rivet head height was measured after insertion using a handheld digital indicator with a standardized collar fixture and are presented in Table 1
Once joined, all specimens were subject to a heat treatment of 180[degrees]C for 30 minutes to simulate an automotive paint shop process. The paint bake process aged the AA6111 to stabilize the alloy and prevent further uncontrolled aging effects on test results.
The quasi-static lap shear failure load of the SPR joints was tested on an Instron electromechanical test frame at cross-head speed of 2 mm/min. Load-controlled tension-tension (R=0.1) fatigue tests were conducted on the SPR joints using a servo-hydraulic MTS test frame at a frequency of 20 Hz. Doublers with a thickness of 2.5 mm were used during fatigue tests to maintain the alignment of the specimen.
Metallurgical analysis was conducted on tested and untested specimens by sectioning along the centerline of the rivet parallel to the loading direction. Sectioned specimens were cold mounted in epoxy and subjected to standard metallurgical techniques. Fine polishing was done with 0.5[micro]m colloidal silica. Macrographs were taken with a Zeiss digital optical microscope. Microhardness measurements were performed on an AMH43 automated hardness indenter. A load of 300 g was applied with a dwell time of 13 seconds. Indentations were made with a spacing of 200 [micro]m to avoid inaccurate readings due to plasticity induced by a nearby indentation. Fractography was conducted on tested specimens using a FEI Nova Nano 650 scanning electron microscope (SEM).
RESULTS AND DISCUSSION
Cross-sections were taken through the centerline of untested joints and subjected to metallurgical preparation as shown in Figure 2a and 2b. Flush rivet head heights, while nominally at zero protrusion from the top surface of the CFRP, resulted in deeper local penetration due to pile-up of the carbon fiber composite from the rivet insertion. As a result, the rivet head damages the composite and delamination is observed between the top ply and the remainder of the sheet, as shown in Figure 2c in the indicated region. While fiber breakage and delamination were also observed in CFRP to aluminum SPR joints produced with proud rivet head height as seen in Figure 2d, the extent of damage was less as compared to SPR joints produced with a flush rivet head height.
Monotonic Lap Shear Failure Load
The quasi-static lap shear load-displacement curves of the CFRP to aluminum and aluminum to aluminum SPR joints are presented in Figure 3 for tests conducted in monotonic loading.
The SPR joints in similar sheets of AA6111 produced the highest overall lap shear failure load of 6231 N. For the joints in CFRP to AA6111, the specimens produced with proud rivet exhibited higher lap shear failure load of 4010 N, while those produced with a flush rivet head height exhibited the lowest lap shear failure load of 2800 N. For all joints considered, failure occurred under monotonic loading by rivet tail pullout in similar AA6111 SPR joints and by head pull-through in both of the CFRP to AA6111 SPR joints as shown in Figure 4.
These findings are consistent with a prior study and suggest that deeper local penetration of the rivet head can result in delamination and fiber breakage in a continuous fiber composite, thereby leading to lower monotonic performance of the SPR joints . This clearly indicates that the damage induced to the laminate by the punched hole at the rivet head plays a crucial role in controlling the failure load of SPR joints with a continuous fiber CFRP top sheet in monotonic loading. Hence, even for the small change in rivet head protrusion considered in this study (+0.29 mm and 0 mm from the top CFRP sheet surface for proud and flush head heights, respectively) could lead to variance in lap shear failure load of the SPR joints by about 1200 N.
Fatigue Test Results
The results of fatigue tests conducted on the CFRP to AA6111 and AA6111 to AA6111 lap shear SPR joints are presented in Figure 5. Each point represents one specimen, and the lines indicate the trends of the data for the three specimen types. All tests presented here were conducted with a load ratio of R=0.1. Tests conducted on SPR joints in similar sheets of AA6111 resulted in the shortest fatigue life. Despite their lower monotonic failure load, the CFRP to AA6111 SPR lap shear joints produced with proud rivet head height exhibited superior fatigue life compared to CFRP to AA6111 SPR joints produced with flush rivet head height. However, for maximum loads above 2500 N, it appears the rivet head height has little to no impact on the fatigue life of SPR joints with a CFRP top sheet.
Furthermore, the fatigue lives of the three specimen types are indistinguishable above 2500 N. Thus, at lower loads, the impact of rivet head height on the fatigue life is more appreciable. Interestingly, irrespective of fatigue life, stack configuration and rivet head height, all the SPR lap shear joints tested at less than 80% of the monotonic failure load in the current study failed in the bottom sheet of AA6111, as shown in Figure 6. Thus, the differences in monotonic failure loads did not represent the differences in fatigue life because the failure modes were different for the two loading conditions (rivet pullout and head pull-through in monotonic loading).
When the maximum load exceeded 80% of the monotonic failure load with R=0.1, all three lap shear specimen types failed due to rivet pullout in less than 500 cycles. This is due to the higher bearing pressure of the rivet at the interface with the CFRP sheet inducing more severe damage to the composite, which can loosen the locking mechanism between the rivet and the CFRP sheet, as discussed in . However, for the range of loads considered at less than 80% of the monotonic failure load in the current study, no fatigue crack growth or failure in the CFRP laminate was observed under cyclic loading. Since it is apparent that two distinct types of failure can occur which may depend on the mode of loading, further testing will be conducted with cross-tension specimens under monotonic and cyclic loading. This specimen geometry will be used as a surrogate for coach peel specimens, which cannot be fabricated from flat plaques of CFRP composites.
For clarity, a representative cross-section of the untested SPR joint in CFRP to aluminum is shown in Figure 7 and indicates the different fracture regions. The crack initiation point, secondary fretting region, fast crack/ductile fracture region are all indicated in the figure. The bold red arrows indicate the crack propagation direction along the width direction of the aluminum sheet.
A representative fracture surface of the AA6111 bottom sheet is shown in Figure 8.
Similar surfaces were observed for all three specimen types. Figure 8a shows a macrograph of the fractured aluminum surface with the top sheet and rivet removed. Figure 8b is a low magnification SEM image of the one-half of the fracture surface. A high magnification SEM image in Figure 8c of the region R1 shows the crack initiation point, and Figure 8d of the region R2 shows the distinctive crack propagation region and fast fracture region. Secondary fretting was also observed closer to the upper surface of the bottom aluminum sheet, and is shown in Figure 8e. The fretting is clearly located in the path of the propagating primary fatigue crack, and is not related to the initiation of the crack. Thus, the role of the secondary fretting on fatigue crack propagation in these SPR joints, if any, still needs to be investigated. The fracture surface analysis of the bottom aluminum sheet indicates the dominant fatigue crack initiated in the cold worked, plastically deformed region close to the rivet-aluminum sheet interlock region (region R1) and propagated along the width of the sheet. During the SPR process, the bottom aluminum sheet undergoes plastic deformation as the rivet fills the stack into the die; concurrently the rivet legs flare into the bottom substrate, creating mechanical interlock. The extent of rivet leg flare and die fill are affected by the rivet penetration into the stack . Although cold work increases the material strength of the aluminum alloy, it also introduces a higher stress concentration . Higher stress concentration may have contributed to the initiation of the dominant fatigue crack. To know the extent of cold working, microhardness map of the local area were acquired. Microhardness measurements were conducted in the cold-worked region of the bottom sheet of AA6111 for each of the three joint types, and the processed maps are presented in Figure 9. The color scale is the same for the all the images. Comparing the joints in CFRP to AA6111 for a rivet head that is proud in Figure 9a, and flush in Figure 9b, it is clear that the flush rivet head height resulted in higher degree of cold work. This could result in a greater stress concentration and reduce the fatigue life of the joint.
The microhardness mapping of aluminum to aluminum SPR joints in Figure 10c showed a degree of cold working similar to the flush CFRP to AA6111 joint close to the rivet. However, in the die area farther from the rivet interface, the extent of cold working appears to be greater. The difference in cold working may be due to slight differences in the degree of springback between the two stacks, or the different constraint applied by the top sheet. The difference in hardness values were also reflected in the microstructure of the cold worked region which were characterized by presence of long, elongated grains. The extent of cold working qualitatively corresponds to fatigue life, in that the joints produced in two sheets of AA6111 had the lowest fatigue life of the tested joints, while the CFRP to AA6111joints with a proud rivet head experienced the longest fatigue life.
FATIGUE LIFE CORRELATION
Structural stress concepts are widely used in fatigue life prediction of spot welds in vehicle structure. In the present study, an attempt is made to correlate the fatigue life of SPR joints in dissimilar CFRP to AA6111, and similar sheets of AA6111, using an existing structural stress model. This structural stress model was developed by Rupp et al.  to correlate the fatigue life of resistance spot welds in automotive body structures. The method allows the determination of the critical stresses in a welded joint based on only a few geometrical inputs of the joint.
As shown conceptually in Figure 10, the structural stress model utilizes two sets of formulae to calculate the critical stress based on the observed failure mode, which include a crack in the sheet, or a crack in the weld nugget. Since all of the specimens in this present study failed in the bottom AA6111 sheet, the appropriate formulae set are presented and are modified accordingly.
The equivalent stresses for sheet failure is calculated using a superposition of formulas for a plate subjected to central loading:
[[sigma].sub.ss] = [[sigma].sub.max]([F.sub.x]) cos[theta] - [[sigma].sub.max]([F.sub.y])sin[theta] + [[sigma].sub.max]([F.sub.z])- [[sigma].sub.max]([M.sub.x])sin[theta] - [[sigma].sub.max]([M.sub.y])COS[theta] (1)
Where [[sigma].sub.max] ([F.sub.x]) = [F.sub.x] /[pi]dt
[[sigma].sub.max] ([F.sub.y]) = [F.sub.y] /[pi]dt
[[sigma].sub.max] ([F.sub.z]) = [kappa]([1.774 [F.sub.z]/[t.sup.2]) for [F.sub.z] > 0
[[sigma].sub.max]([M.sub.x]) = [kappa]((1.872 [M.sub.x]/(d[t.sup.2]))
[[sigma].sub.max]([M.sub.y]) = [kappa]((1.872 [M.sub.y]/(d[t.sup.2]))
Where d is the diameter of the rivet at the failure region, t is the thickness of the sheet that failed/cracked, [kappa]= 0.6[square root of (t)], Mx, My, Fx, Fy, are bending moments and maximum applied forces in the sheet. The forces and moments in equation 1 are calculated using a simple free body diagram. Since failure occurs in the bottom aluminum sheet at a fairly consistent location likely due to the stress concentration introduced by the rivet interlock in the lower sheet, the diameter of the rivet at the failure location was calculated for each of the SPR joints, as shown in a representative SPR joint in Figure 11. The diameters, d, for each SPR are presented in Table 2, and are different for each material stack-up and rivet head height, the latter corresponding to a different degree of leg engagement and consequent flaring in the rivet shank.
Using the above inputs and equation, structural stresses were calculated for each of the three SPR joints. It has to be noted, the diameters mentioned in Table 2 are average of two as-received specimens from each set of SPR joints. Finally, the calculated structural stresses of the SPR joints are replotted with the experimentally-observed fatigue lives, and presented in Figure 12.
From Figure 12, it is observed that the fatigue life from the test results versus the calculated structural stresses gives a good correlation. Irrespective of material stacking and rivet head height, the fatigue life of the SPR joints fall in one single master curve. Thus, the differences in fatigue life resulting from proud and flush head heights considered in the present study are well-captured by the structural stress model for the given lap shear joint configuration. Overall, the Rupp's structural stress model commonly used for resistance spot welded joints can be adapted to predict the fatigue life of self-piercing rivet joints.
Failure load and fatigue properties of lap shear SPR joints in similar sheets of AA6111 and dissimilar sheets of continuous braided carbon fiber reinforced polymer composite and AA6111 were evaluated in this study. Based on the testing results, the following conclusions can be drawn:
1. Self-piercing rivet can be used to join carbon fiber reinforced plastic composite to aluminum alloys. A maximum average lap shear failure load of 4010 N was achieved in CFRP to AA6111 SPR joints.
2. The rivet head piercing, that is a flush head height versus a proud head height plays a dominant role in controlling the tensile and fatigue properties of the CFRP to aluminum SPR joints. In quasi-static lap shear monotonic tests, CFRP to AA6111 SPR joints produced with proud rivet head height exhibited a maximum load that was 1200 N greater than those produced with a flush rivet head height. A flush rivet head height resulted in more severe local damage of the CFRP sheet, as seen in greater material pile-up at the rivet and delamination. This significantly reduced the lap shear failure load. In both the SPR joints, the mode of failure was due to rivet pullout.
3. In fatigue tests, the CFRP to AA6111 SPR joints with a proud rivet head height exhibited higher fatigue life compared to those with a flush rivet head height. Overall, the CFRP to AA6111 SPR joints exhibited better fatigue life compared to SPR joints in similar sheets of AA6111. The dominant mode of failure in all the SPR joints was fatigue crack growth through the sheet width. The fatigue crack in all three joint types was found to initiate in the plastically deformed region close to the interface of the rivet and aluminum sheet. Microhardness maps of the plastically deformed region indicate a high degree of cold work. The extent of cold work qualitatively correlated to the fatigue life of the three joint types.
4. A mathematical model based on Rupp's structural stress approach was adopted to develop a master curve. Using the diameter of the rivet at the point of failure for each joint type, it is possible to develop a reliable master curve which provides fatigue life of the SPR joints at different loads.
(1.) Eusebi, E., "Composite intensive vehicles: past present and future," Struct. Mater. Challenges next Gener. Veh. Washington, DC US Dep. Commer., 1995.
(2.) Gjostein, N.A., "Technology needs beyond PNGV," Basic Needs Veh. Futur. New Orleans, Louisiana (5 January 1995), 1995.
(3.) Cai, W., Wang, P.C., and Yang, W., "Assembly dimensional prediction for self-piercing riveted aluminum panels," Int. J. Mach. Tools Manuf. 45:695-704, 2005, doi:10.1016/j.ijmachtools.2004.09.023.
(4.) Mortimer, J., "Jaguar uses X350 car to pioneer use of self-piercing rivets," Ind. Robot An Int. J. 28:192-198, 2001, doi:10.1108/01439910110389344.
(5.) Barnes, T.A. and Pashby, I.R., "Joining techniques for aluminum spaceframes used in automobiles. Part II -adhesive bonding and mechanical fasteners," J. Mater. Process. Technol. 99:72-79, 2000, doi:10.1016/S0924-0136(99)00361-1.
(6.) Fu, M. and Mallick, P.K., "Fatigue of self-piercing riveted joints in aluminum alloy 6111," Int. J. Fatigue 25(3):183-189, 2003, doi:10.1016/S0142-1123(02)00115-9.
(7.) Lin, P. and Su, Z., "Fatigue Behavior of Self-Piercing Rivets and Clinch Joints in Lap-Shear Specimens of Aluminum Sheets CLINCH JOINTS OF 6111-T4," Int. J. Fatigue 72:293-298, 2013, doi:10.4271/2013-01-1024.
(8.) Miyashita, Y., Jack Teow, Y.C., Karasawa, T., Aoyagi, N., Otsuka, Y., and Mutoh, Y., "Strength of adhesive aided SPR joint for AM50 magnesium alloy sheets," Procedia Eng. 10:2532-2537, 2011, doi:10.1016/j.proeng.2011.04.417.
(9.) Abe, Y., Kato, T., and Mori, K., "Self-piercing riveting of high tensile strength steel and aluminium alloy sheets using conventional rivet and die," J. Mater. Process. Technol. 209:3914-3922, 2009, doi:10.1016j.jmatprotec.2008.09.007.
(10.) Mori, K., Kato, T., Abe, Y., and Ravshanbek, Y., "Plastic joining of ultra high strength steel and aluminium alloy sheets by self piercing rivet\r\n," CIRP Ann. - Manuf. Technol. 55(2):1-4, 2006, doi:10.1016/S0007-8506(07)60417-X.
(11.) Franco, G. Di, Fratini, L., Pasta, A., and Ruisi, V.F., "On the self-piercing riveting of aluminium blanks and carbon fibre composite panels," Int. J. Mater. Form. 6(1):137-144, 2013, doi:10.1007/s12289-011-1067-2.
(12.) Franco, G. Di, Fratini, L., and Pasta, a., "Analysis of the mechanical performance of hybrid (SPR/bonded) single-lap joints between CFRP panels and aluminum blanks," Int. J. Adhes. Adhes. 41:24-32, 2013, doi:10.1016/j.ijadhadh.2012.10.008.
(13.) Zhang, J. and Yang, S., "Self-piercing riveting of aluminum alloy and thermoplastic composites," J. Compos. Mater., 2014, doi:10.1177/0021998314535456.
(14.) Kroll, L., Mueller, S., Mauermann, R., Gruetzner, R., Technology, F., Structures, L., and Gmbh, E., "Strength of Self-Piercing Riveted Joints for Cfrp / Aluminium Sheets," 18 Th Int. Conf. Compos. Mater. 1-6, 2011.
(15.) Kang, J., Rao, H., Zhang, R., Avery, K., and Su, X., "Tensile and fatigue behaviour of self-piercing rivets of CFRP to aluminium for automotive application," IOP Conf. Ser. Mater. Sci. Eng. 137(1):012-025, 2016, doi:10.1088/1757-899X/137/1/012025.
(16.) Han, L., Chrysanthou, a., Young, K.W., and O'Sullivan, J.M., "Characterization of fretting fatigue in self-piercing riveted aluminium alloy sheets," Fatigue Fract. Eng. Mater. Struct. 29:646-654, 2006, doi:10.1111/j.1460-2695.2006.01045.x.
(17.) Han, L. and Chrysanthou, A., "Evaluation of quality and behaviour of self-piercing riveted aluminium to high strength low alloy sheets with different surface coatings,"Mater. Des. 29:458-468, 2008, doi:10.1016/j.matdes.2006.12.020.
(18.) Pickin, C.G., Young, K., and Tuersley, I., "Joining of lightweight sandwich sheets to aluminium using self-pierce riveting," Mater. Des. 28:2361-2365, 2007, doi:10.1016/j.matdes.2006.08.003.
(19.) Franco, G. Di, Fratini, L., and Pasta, A., "Influence of the distance between rivets in self-piercing riveting bonded joints made of carbon fiber panels and AA2024 blanks," Mater. Des. 35:342-349, 2012, doi:10.1016/j.matdes.2011.09.036.
(20.) Fiore, V., Bella, G. Di, Galtieri, G., Alagna, F., Borsellino, C., and Valenza, A., "Mechanical Behaviour of Spr / Co-Cured Composite To Aluminium Joints," ECCM15-15th European Conference of Composite Materials, Venice, Italy: 24-28, 2012.
(21.) Fratini, L. and Ruisi, V.F., "Self-piercing riveting for aluminium alloys-composites hybrid joints," Int. J. Adv. Manuf. Technol. 43(1-2):61-66, 2009, doi:10.1007/s00170-008-1690-3.
(22.) Rupp, A., Storzel, K., and Grubisic, V., "Computer Aided Dimensioning of Spot-Welded Automotive Structures," SAE Technical Paper 950711, 1995, doi:10.4271/950711.
(23.) Godwin, E.W., Matthews, F.L., "A Review of the Strength of Joints in Fibre Reinforced Plastics - Mechanically-Fastened Joints."," Composites 11(July):150-160, 1980.
This work was supported by the Ford Motor Company. The assistance provided by Dr. Joao Carvalho de Moraes to conduct microhardness measurements is gratefully acknowledged. The authors additionally thank Mr. Mark Deleon and Mr. John Morris for fabricating the joints considered in this study. The authors would also acknowledge the assistance provided by Ms. Renata Zevadil in performing the failure analysis and metallographic preparation.
Harish M. Rao and Jidong Kang
Canmet MATERIALS Technology Laboratory
Garret Huff, Katherine Avery, and Xuming Su
Ford Motor Company
Table 1. Stack configuration and rivet head height Stack configuration Rivet head height (mm) Average (range) CFRP to Al (proud rivet head height) 0.29 (+0.06/-0.07) CFRP to Al (flush rivet head height) 0.00 (+0.10/-0.06) Al to Al (flush rivet head height) -0.01 (+0.07/-0.04) Table 2. Geometrical inputs for each SPR joint. Rivet diameter, d, Stack configuration (mm) CFRP to Al (proud rivet head height) 6.0 CFRP to Al (flush rivet head height) 6.2 Al to Al (flush rivet head height) 6.4 Sheet thickness, t, Stack configuration (mm) CFRP to Al (proud rivet head height) 2.5 CFRP to Al (flush rivet head height) 2.5 Al to Al (flush rivet head height) 2.5
|Printer friendly Cite/link Email Feedback|
|Author:||Rao, Harish M.; Kang, Jidong; Huff, Garret; Avery, Katherine; Su, Xuming|
|Publication:||SAE International Journal of Materials and Manufacturing|
|Date:||May 1, 2017|
|Previous Article:||Finite Element Analyses of Stress Intensity Factor Solutions for Discontinuous Gas Metal Arc Welds under Lap-Shear Loading Conditions.|
|Next Article:||Fatigue Analysis of Swept Friction Stir Clinch Joints between Aluminum and Steel Sheets.|