Development of intelligent sensing unit for micro-electrical discharge machining

Evaluating workpiece quality of electrical discharge machining (EDM) process is challenging users when trying to extract features from the stochastic and time-consuming process. This study proposes an intelligent sensing unit for EDM (ISU-EDM) to extract key machining features for predicting the roughness of workpieces. In machining, the ISU-EDM can synchronously sample the waveforms of discharge current and voltage, and automatically segment the waveforms as sensing data according to tool location and waveform effectiveness. In further, the machining features are extracted from the sensing data by using a genetic-algorithm-based distribution fitting method. The experimental results show that the mean absolute percentage error of prediction accuracy of roughness is less than 15% after applying the extracted features to the automated virtual metrology system.