Soft computing-based hybrid heuristic optimization of wear performance in friction stir-processed composites

Soft computing methods are now powerful aids for optimizing the parameters of material processing and improving wear resistance. This research proposes a hybrid heuristic method for optimizing Friction Stir Processing (FSP) parameters for improved wear properties. This research started by examining the key processing parameters of tool rotation speed, traverse speed, and number of passes using a face-centered central composite design (FCCCD) under response surface methodology (RSM).To determine and optimally wear resistance, an integrated computational approach using RSM, Genetic Algorithm (GA), Artificial Neural Networks (ANN), and Adaptive Neuro-Fuzzy Inference System (ANFIS) was used. The GA-ANN model obtained an accuracy of 95.48%, while the GA-ANFIS model yielded an accuracy of 92.9%, confirming that the two models proved to be appropriate predictive models of wear behavior. Microstructural analysis using Field Emission Scanning Electron Microscopy (FESEM) showed that when using the optimized conditions for FSP, the reinforcement particles dispersed uniformly, reduced the formation of wear debris, and improved wear resistance and tribological properties. The morphology of the worn surfaces demonstrated a change from severe delamination and deep plowing for the lower performing surface to a refined abrasive to explain the reduction in material loss and improved performance. Furthermore, three-dimensional response surface plots illustrated the significant role of rotational speed and the number of passes on wear resistance, underlining the importance of parameter selection for superior tribological performance.