Integrated Methodology of Soft Computing for Process Modeling and Optimization of Duplex Turn Cutting of Titanium Alloy

Abstract

Duplex turning (DT) is a novel concept of metal cutting where two tools are employed to cut the objects in lieu of single tool. It shows many benefits over conventional turning in terms of superior dynamic balancing, lower cutting forces and tool wears, better surface finish, reduction in vibration with additional support for workpiece. It is a complex method and the resulting experimental analysis becomes difficult and expensive. In such conditions, modeling techniques show their potential for parametric study, prediction of data for optimization and selection of optimal condition. Generally, soft computing-based Artificial Neural Network (ANN) is applied for modeling and prediction for complicated processes while Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) shows their potential for optimization of complex problems over Genetic Algorithm. Therefore, ANN and NSGA-II techniques are employed for modeling and optimization of DT process to minimize the surface roughness and cutting forces (primary and secondary). Finally, results reflect that ANN efficiently predicts the responses at different input combinations within training data set with absolute percentage errors as 2.55% for roughness, while 3.05% and 3.14% for cutting forces (primary and secondary), respectively. In the same way, optimized results also found within the range of acceptability with percentage errors as 2.57% for roughness, while 3.25% and 3.15% for primary and secondary forces, respectively.