Predictive Modelling and Optimization of Surface Roughness in SPIF of Ti-6Al-4V Alloy Using ANN for Marine Applications

Abstract

Being a flexible and efficient process of manufacturing complex sheet metal products, the Surface Roughness of SPIF process is one of the problems to be solved. This paper aims at analyzing the mean surface roughness Ra when performing the SPIF process using CNC milling machine. Surface roughness (Ra) affects the durability of the product and leads to its premature breakdown; therefore, the optimization of the process and predicting its results is important. The titanium alloy (Ti-6Al-4V) is chosen for this experiment due to its high resistance to corrosion and excellent specific strength. The Taguchi method was employed to investigate the spindle speed, tool diameter, tool feed, and incremental step depth for the experiments. ANOVA results show that incremental step depth is the dominant parameter affecting Ra, followed by tool feed and tool diameter with moderate influence, while spindle speed has a negligible effect. Taguchi optimization confirms the optimal parameters effectively minimize Ra, showing good consistency between predicted and experimental data. Furthermore, regression and ANN models were developed and validated using Minitab® and MATLAB®, respectively, showing strong agreement with experimental data and high predictive accuracy. The regression model exhibited strong statistical performance, while the ANN model achieved high correlation with low error.