Online Workpiece Height Estimation for Reciprocated Traveling Wire EDM Based on Support Vector Machine

Abstract

The reciprocated traveling wire EDM is a unique type of wire EDM machine, in which the wire electrode is not discarded after passing the discharge gap but being re-winded on the wire drum instead. It is of specific features such as ultra-thick workpiece cutting and reuse of wire electrode. While cutting variable height workpieces with constant machining parameters, this type of WEDM faces the same challenges as western style WEDM, either the risk of wire breakage or the low cutting speed. Due to the intermittent discharge process and the varying machining characteristics, online estimation of workpiece height becomes even more difficult for this type of WEDM.This paper proposes a method of online workpiece height estimation based on support vector machine (SVM), an effective machine learning method. The inputs of the SVM model are effective discharge frequency, pulse interval, programmed feed rate and actual feed rate and the output is the estimated height. The algorithm is integrated in a newly developed computer numerical control (CNC) system for reciprocated traveling WEDM, with a sampling circuit collecting current and voltage signals from the discharge gap and an adaptive control unit which adjusts the machining parameters according to the workpiece height estimation.The data for training the SVM were produced by cutting stair-shaped workpieces to build the SVM model. Then verification was carried out by cutting through workpieces with variable height cross sections. The results demonstrated the effectiveness of the proposed method. The estimation error was less than 2mm and the machining time was reduced by more than 30%.