A real-time predictive control method of discharge state for micro-EDM based on calamities grey prediction theory

In micro-electrical discharge machining (micro-EDM) process, the property of high-frequency discharge and complex interference causes the discharge state to vary fast, and conventional control methods of micro-EDM have the hysteresis characteristic, which make it very difficult to conduct efficient and stable control. Firstly, in this paper, based on calamities grey prediction theory, a new method for discharge state prediction is proposed, where the voltage and current signal are replaced with the output velocity of conventional control system as the direct predictive target, and the negative velocity (electrode pullback) is regarded as a calamity value. Owing to avoidance of processing voltage and current signals with nonstationarity, nonlinearity, and internal coupling characteristics, the proposed discharge state prediction method, which predicts that whether the future processing cycles are in short-circuit discharge state, exhibits good real-time performance. Thereafter, by virtue of multi-thread technology, the proposed discharge state-predictive method is added to a conventional control system, and a discharge state real-time predictive control system of micro-EDM is established in this paper. Finally, experiments on machining micro-holes and micro-grooves are conducted using the proposed predictive control method and fuzzy control method. Experiment results indicate that the proposed predictive control method has stable and good real-time performance, and it can effectively reduce electrode pullbacks and improve the processing efficiency and processing quality in both micro-hole and micro-groove EDM.