Failure detection and control for wire EDM process using multiple sensors

Abstract

Unstable machining conditions during wire EDM process can lead to process failures, which affects the efficiency and sustainability of process. The study aims to develop a sensor-based failure prediction and process control system. The monitoring system consisting of high sampling rate differential and current probes extracts voltage and current signals during spark machining. Relevant discharge characteristics like pulse proportions, pulse frequency, and discharge energy are extracted from the pulse train data. The proposed process control algorithm works in three stages: failure prediction, failure severity assessment, and process control. Failure conditions considered are wire breakage and spark absence, which are predicted based on the extracted discharge characteristics. Severity of failure is judged based on the spark discharge energy. The proposed process control algorithm retunes the process parameters by adjusting pulse on time, pulse off time, and servo voltage, based on the type of failure and its severity. The methodology is successful in preventing the potential failure situation by restoring the machining stability. The capability of the model is demonstrated by conducting confirmation experiments. Microstructural comparison of machined surfaces and worn wire surfaces also confirms the effectiveness of the proposed strategy to ensure failure free machining with better surface integrity.