ABSTRACTCast alloys find suitable employability in industrial applications favoring automotive and aerospace sectors owing to their high specific strength, low porosity, excellent fluidity, and machinability. The present study focuses on friction stir welding of cast aluminum alloys (AA356 and AA2014) with a varied range of process parameters, namely, tool pin shape (cylinder, cone, square, and threaded cylinder), tool speed (1800–2100 rpm), and tool feed (5–20 mm min−1). The novelty of the present work spotlights/features the implementation of L16 orthogonal design with the analysis of variance, grey relational analysis, fuzzy logic system, and artificial neural network approaches to contemplate the weld quality of cast joints. The microstructural analysis and grain size estimation is comprehended in the study. Regression models have been developed based on the analysis of variance. The prime factors favoring the ultimate tensile strength and microhardness were 2100 rpm, 10 mm min−1, and threaded cylinder shape. The findings from the analysis of variance and grey relational analysis are in concurrence with each other. The results of fuzzy logic showcased a substantial improvement in terms of grey relational grade. The results from the artificial neural network deliver rational evidence for the analysis of variance conducted on the selected tool pin shapes.
Read full abstract