Soft computing model for analysing the effect of friction stir processing parameters on the intergranular corrosion susceptibility of aluminium alloy AA5083

Abstract

Aluminium alloy AA5083 is prone to intergranular corrosion in marine environments. In an attempt to reduce the intergranular corrosion, AA5083 was subjected to friction stir processing (FSP). The FSP experimental trials were conducted as per face-centered central composite design with three levels of variation in FSP process parameters viz. tool rotation speed (TRS), tool traverse speed (TTS) and tool shoulder diameter (SD). Intergranular corrosion susceptibility of the processed specimens was assessed by performing nitric acid mass loss test. The mass loss of the specimens was correlated with the intergranular corrosion susceptibility as per the standard ASTM G67-13. The experimental results indicate that FSP had significantly reduced the intergranular corrosion susceptibility of the AA5083 alloy. Soft computing techniques namely Artificial Neural Network, Mamdani Fuzzy system, and Sugeno Fuzzy system were used to predict the intergranular corrosion (IGC) susceptibility (mass loss) of the friction stir processed specimens. Among the developed models, Sugeno fuzzy system displayed minimum percentage error in prediction. So Sugeno fuzzy system was used to analyze the effect of friction stir processing process parameters on the IGC of the processed specimens. The results suggest that stir processing of AA5083 at a TRS of 1300 rpm, TTS of 60 mm/min and SD of 21 mm would make the alloy least susceptible to intergranular corrosion.