The study of machine learning for wire rupture prediction in WEDM

During wire electrical discharge machining (WEDM), rupture may deteriorate workpieces’ machined surfaces and increase the processing time. However, only a few referenced papers focused on during past decades because of its complexity. In this research, machine learning (ML) technique was applied to analyze relationship between manufacturing parameters chance rupture. Three parameters, including gap voltage (GV), feed rate (FR), water resistance (WR), were considered as training features, total 298 sets used train an artificial neural network (ANN). The results show that prediction accuracy for 10 s in advance is above 85%. This research developed new method apply real-time predict faster, more accurate than prior research. Besides, extendable future measured data when usable sensor are increasing.