Surface Roughness Prediction in EDM: Integration of Experimental Design and Machine Learning

This chapter delves into the intricate world of Electrical Discharge Machining (EDM) and its impact on surface roughness, particularly in C45 steel. The study focuses on three key parameters: current intensity (I), pulse-on time (Ton), and pulse-off time (Toff). Through a meticulously designed experiment, it reveals how increasing current intensity and pulse-on time generally degrades surface finish, while the effect of pulse-off time is more nuanced and depends on interactions with other parameters. The chapter also explores the application of machine learning models, specifically Support Vector Regression (SVR) and linear regression, to predict surface roughness. The SVR model demonstrates a significant improvement over linear regression, achieving an R² of 79%, highlighting the complex non-linear relationships in EDM processes. The findings provide valuable insights for optimizing EDM processes in industrial applications, particularly where surface integrity is critical. The chapter concludes by suggesting future research directions, including multi-objective optimization approaches and real-time control systems.