Material Characterization and Modeling of Friction Stir Spot Welds in a Magnesium AZ31 Alloy
The fatigue behavior of friction stir spot welds in magnesium AZ31 alloy is experimentally investigated and modeled. The friction stir spot welds employed in this study are representative of preliminary welds made in developing the joining process for potential use in automobile manufacturing. Load control cyclic tests were conducted on single weld lap-shear coupons to determine fatigue life properties. Optical fractography of the failed fatigue coupons revealed that fatigue cracks initiated from the interfacial “hook” and eventually failed by either nugget pullout or full width separation, depending on the cyclic load amplitude. The failure modes of the magnesium AZ31 alloy were similar to the aluminum alloys of comparable friction stir spot welds. To predict the fatigue life of the lap-joint coupons, a crack growth modeling approach based on a kinked crack stress intensity solution was used. The fatigue model predictions compared well to the experimental fatigue life results, despite an approximate stress intensity factor solution for this weld geometry. The experiments and modeling conducted in this study suggest that the size of the interfacial hook, which comes about from the speed, depth of plunge, dwell time, and tool configuration of the friction stir spot weld process, is a major contributor to the fatigue life of the joint.
With the increased interest in achieving improved fuel economy in automobiles, the need for lightweight alloys is essential. While only marginally employed in current automobile designs, recent advances have produced promising magnesium alloys that are now being considered for structural components. As with any newly employed alloy, research is needed to determine the reliability and performance of joints made from potential joining techniques. With the friction stir spot welds (FSSW) technology advancing, the need to evaluate the fatigue performance of this joining technique on other lightweight alloys, i.e., magnesium alloys, is the basis for this study. Limited published literature exists on the fatigue performance of FSSW coupons of magnesium alloys  . While Mallick and Agarwal  quantified the fatigue performance of the FSSW in a magnesium AM60 alloy, characterization of the failure mechanisms was limited and no modeling comparison was presented. As such, this is the first research of its kind to quantify the fatigue performance and failure mechanisms of the FSSW in a magnesium alloy and evaluate the fatigue predictions of an engineering model.
The FSSW tool, shown in Fig. 2, was made from standard tool steel (H13) having a shoulder with a diameter of 12 mm, pin length of 3.2 mm, and left hand threads (M5). The welding process parameters were: tool rotation speed of 1000 rpm, tool plunge speed of 20 mm/min, shoulder plunge depth of 0.5 mm, and a dwell time of 2.5 s.
A total of three FSSW coupons were monotonically tested to failure and the average ultimate strength was approximately 3.7 kN. Based on the ultimate strength of the FSSW coupons, fatigue testing was performed at maximum load levels ranging from 1-3 kN, as shown in Table 1. It is important to note that the fatigue crack always grew from the primary hook and no crack growth was observed from the secondary hook. Figure 11 shows the fatigue crack path for each of the load levels tested.
Regarding the fatigue mechanisms for the magnesium AZ31 alloy, the cross section views shown in Fig. 11 revealed two types of fatigue failure, resulting in complete separation. The nugget pullout failure observed for the two maximum load levels tested (2.5 kN and 3.0 kN, R=0) likely occurred because as the crack (denoted as primary crack (a) in Fig. 11) propagated circumferentially around the nugget, the shear/tensile stress in the remaining net area of the nugget increased with each advancement of the crack front.
The use of a crack growth model to predict the fatigue life of the welded lap-joint components is not unique. The use of the crack growth approach employed here requires some simplifying assumptions. The major assumptions employed in this paper follow similar assumptions listed in part by Newman and Dowling  .
1. A three-dimensional crack behaves like a planar crack and can be treated like a two-dimensional crack. 2. The fatigue crack always initiates from the same location. 3. The fatigue life consists mainly of crack growth. 4. The stress intensity solution based on a RSW is a reasonable approximation for a FSSW. 5. The modes I and II stress intensity factors remain constant and are independent of crack length. 6. An equivalent mode I stress intensity solution is valid for mixed mode loading.
The crack growth model compared well, if not slightly conservative, to the FSSW fatigue life results for the three R-values used, as seen in Fig. 13. In addition, the crack growth rate predicted by the model was within an order of magnitude to the crack growth rate estimated from the striation spacing for the R=0 tests.
Based on the experimental and modeling work conducted in this study, the main influencing parameter affecting the fatigue performance of the FSSW made of magnesium AZ31 alloy is the height of the interfacial hook.
- ↑  Mallick, P. K., and Agarwal, L., 2009, “Fatigue of Spot Friction Welded Joints of Mg-Mg, Al-Al and Al-Mg Alloys,” Society of Automotive Engineers, War- rendale, PA, SAE Technical Paper No. 2009-01-0024.
- ↑  Newman, J. A., and Dowling, N. E., 1998, “A Crack Growth Approach to Life Prediction of Spot-Welded Lap Joints,” Fatigue Fract. Eng. Mater. Struct., 21, pp. 1123–1132.
Citation: Fatigue Characterization and Modeling of Friction Stir Spot Welds in Magnesium AZ31 Alloy, Jordon, J.B., M.F. Horstemeyer, S.R. Daniewicz, H. Badarinarayan, and J. Grantham, Journal of Engineering Materials and Technology, 132, (2010)