Application of Machine Learning Techniques for Multi Objective Optimization of Response Variables in Wire Cut Electro Discharge Machining Operation

Abstract

Wire Cut Electrical Discharge Machining (WEDM) is a non-conventional thermal machining process which is capable of accurately machine alloys having high hardness or part having complex shapes that are very difficult to be machined by the conventional machining processes. The WEDM finds applications in automobiles, aero–space, medical instruments, tool and die industries, etc. The input parameters considered for WEDM are pulse on time, pulse off time, flushing pressure, servo voltage, wire feed rate and wire tension. Performance of WEDM is mainly assessed by output variables such as, material removal rate (MRR), kerf width (Kw) and surface roughness (Ra) of the work piece being machined. Looking at the need of a suitable optimization model, the present work explores the feasibility of machine learning concepts to predict optimum surface roughness and kerf width simultaneously by making use of experimental data available in the literature for machining of Hastelloy C– 276 using WEDM. In most of the literatures, single objective optimization has been carried out for predicting optimum cutting parameters for WEDM. Hence, the present work presents a methodology that makes use of a machine learning algorithm namely, gradient descent method as an optimization technique to optimize both surface roughness and kerf width simultaneously (multi objective optimization) and compare the results with the existing literatures. It was observed that the input parameters such as pulse on time, pulse off time, and peak current have significant effect on both surface roughness and kerf width. The gradient descent method was successfully used for predicting the optimum values of response variables.

A Comprehensive Review of Machine Learning in Multi-objective Optimization

“…Gradient descent is applied in [60,61] on MOO in the domains of discharging process and bioprinting, respectively. Bayesian optimization is used for approximating the Pareto Front set in [62].…”

Surface roughness prediction of AISI D2 tool steel during powder mixed EDM using supervised machine learning

“… References Year Machined materials Input parameters Response parameters Prediction method 23 2022 NiTi alloys, NiCu alloys and BCu alloys Ton, Toff, Gap current and gap voltage MRR RF, D.T., Gradient Boosting ANN 24 2020 Aluminum Voltage, Ton, Wire feed, dielectric pressure S.R Support vector method (SVM), Extreme learning machine, Weighted Extreme learning machine 25 2022 Shape memory alloys (SMA) Nitinol rods Ton, Toff and current S.R AlexNet, KNN, MNB and DenseNet 26 2022 Memory alloy of Cu-based shape Peak current (Ip), Ton, gap voltage and Toff D.D. and TWR GA and TLBO techniques 27 2021 Inconel 718 wire feed rate, Ip, Ton, Toff and servo voltage SR ANN, SVM 28 2021 EN31 tool steel Ton, Toff, Ip, Vg, flushing pressure (P) Tool shape prediction and S .R DT, R.F., linear model and ANN 29 2020 Hastelloy C-276 Wire tension, flushing pressure, Ton, Toff, servo voltage and wire feed rate Kerf width and S.R Gradient descent method 30 2022 ...…”

Machine Learning Based Predictive Modeling of Electrical Discharge Machining of Cryo-Treated NiTi, NiCu and BeCu Alloys

“…On other occasions regarding using ML approaches, several techniques were used to investigate the process output, such as MRR and predictive model of EDM operations. Shukla and Priyadarshini [13] successfully used a gradient descent method as ML algorithm simultaneously optimise surface roughness and kerf width. It was seen that pulse on and off times and peak current significantly affect the roughness and width of the kerf.…”