Expert system for wire cutting EDM, based on pulse classification and thermal modeling

Abstract

Although wire EDM has become a fully competitive machining technique during the last decade, the process performance is limited by some typical restrictions: the risk of wire rupture and the bending of the wire. Wire rupture drastically reduces the overall efficiency of the EDM operation and wire deflection restricts the obtainable geometrical precision. In order to improve machining performance, a multidisciplinary expert system was developed. It mainly consists of two parts: an observation procedure and a decision procedure. The observation of the wire EDM process is carried out by means of a powerful electronic device, called the EDM pulse discriminating system, which is used to detect and classify the occurring electric pulse types times. This device delivers “on-line” a “chaos” of information of the process. In the decision phase, the incoming data are analyzed and patterns are detected by means of a computer, and eventually, the parameter settings of the EDM machine are changed in order to obtain more optimal working conditions. This decision procedure is partially based on the results of a mathematical thermal model of the wire, enabling one to predict the influence of some parameter changes. Further, a technological database is permanently accessible for the control strategy program. As a result, supervision of the machining process is completely automated. Machining speed can be increased without enhancing the risk of wire rupture, because in the case of an increased danger of rupture, appropriate action is taken by the control strategy. The paper gives some results on how the thermal model may predict some thermal overload of the wire. It also describes how frequently observed breakage at sharp corners of the wire path may be dealt with. It further illustrates the actions taken by the expert system when various types of disturbance occur. The consideration of practical experience with a thermal model yields an increased level of machine autonomy ensuring a performance level obtainable only with a skilled operator.