Process Parameter Optimization in Refill Friction Stir Spot Welding of 6061-T6 Aluminum Alloy and AZ31B-H24 Magnesium Alloy Dissimilar Joints

Lightweight multi-material structures that pair aluminum with magnesium are increasingly sought in the transport and aerospace sectors, yet sound dissimilar Al–Mg joints remain difficult to obtain because the two metals readily form hard, brittle Al–Mg intermetallic compounds (IMCs) whose growth along the bond line degrades joint strength. Refill friction stir spot welding (RFSSW) is a keyhole-free, solid-state process well suited to limiting this reaction, but a validated, quantitative process window for the AA6061-T6/AZ31B-H24 couple is still lacking. This study addresses that gap by establishing quantitative correlations between the governing RFSSW parameters, namely tool rotational speed (1800–2400 rpm), sleeve penetration depth (2.2–2.6 mm) and welding time (4–6 s), and the resulting thermal field, joint integrity and mechanical performance. A two-dimensional axisymmetric, fully coupled thermomechanical finite element model was developed in Simufact Forming 2025.2 software and validated against embedded K-type thermocouple and infrared thermography measurements, with the predicted and measured peaks agreeing to within about 5% (168 °C at 4 s and 216 °C at 6 s, recorded 5 mm from the weld center). The validated model predicts stir-zone peak temperatures of 320–498 °C across the process window; these remain below the solidus of both base alloys (582 °C for AA6061-T6 and 566 °C for AZ31B-H24), confirming a bulk solid-state process, while at the hottest settings, the interfacial temperature locally exceeds the Al–Mg eutectic range (437–450 °C), allowing transient eutectic liquation confined to the faying surface. Lap-shear testing showed that the peak load rose monotonically with penetration depth and welding time to a maximum of 3.08 kN (2.6 mm/6 s) but that maximum strength and maximum energy absorption were not obtained under the same condition: the best strength–toughness balance was provided by the 2.6 mm/4 s condition (2.98 kN peak load, 3.6 J absorbed energy), and all joints failed by nugget (button) pull-out. The principal contribution of the study is a physically grounded, experimentally validated parameter window demonstrating that an intermediate heat input, rather than the highest, is preferable when both strength and toughness are required, offering practical guidance for lightweight structural applications in transportation and aerospace engineering. In practical terms, the parameter combination recommended for a balanced dissimilar joint is 2.6 mm sleeve penetration with a 4 s welding time (tested at 2400 rpm), which delivered a 2.98 kN peak load together with the largest ductile tail of the series.