Parametric Study, Simulation and Optimization of a Friction Stir Welding

This thesis research implemented an existing thermo mechanical model of friction stir welding process, and studied the surrogate model-based optimization approach to obtain optimal process parameters for the modelled friction stir welding process. As an initial step, the thermo mechanical model developed by Zhu and Chao for friction stir welding of 304L stainless steel was replicated using ANSYS. The developed model was then used to conduct parametric studies to understand the effect of various input parameters like total rate of heat input, welding speed and clamping location on temperature distribution and residual stress in the work piece. With the data from the simulated model, linear and nonlinear surrogate models were constructed using regression analysis to relate the selected input process parameters with response variables. Constrained optimization models were formulated using surrogate models and optimization of process parameters for minimizing cost and maximizing throughput was carried out using improved harmony search algorithm. To handle the constraints, Deb’s parameter-less penalty method was used and implemented in the algorithm. It is learned from this research that: (1) heat input is mainly constrained by the lower bound of the temperature for making good welds; (2) the optimal welding speed must balance the loss of heat input and the gain in productivity; (3) clamping closer to the weld is better than away from the weld in terms of lowering the peak residual stresses. Moreover, the nonlinear surrogate models resulted in a slightly better optimal solution than the linear models when wide temperature range was used. However, for tight temperature constraints, optimization on linear surrogate models produced better results. The implemented improved harmony search algorithm seems not able to converge to the best solution in every run. Nevertheless, the non-converged solution it found was very close to the best.