放電加工之波形判別與模糊控制設計

Abstract

本篇論文在於探討放電加工機的放電間隙電壓波形，並針對不同的波形提 出模糊控制的法則。在波形的判別上面，我們將波形分為八類十八種波形 ，並利用波形的特徵定出不同的特徵參數，以六個特徵的判定建出波形產 生矩陣（pulse generator matrix），再根據波形產生矩陣，建立十八個 不同的布林函數，用來判別不同種類的十八種波形。波形判別完成之後， 本文亦提出預測下一個波形的方法，並進而建立模糊控制法則表，提供一 個模糊控制的雛形以供參考。 The aim of this paper is to discuss gap voltage pulse identification in an EDM process and to propose a fuzzy control rule table. Generally, the discharge waveforms of normal sparks, short circuits, arcs and open circuits can be distinguished easily. Few systems have more advanced features like discharge voltage noise detection facilities, particular arc type detections, etc. It turns out that some process phenomena can not fully be clarified and understood. So, this paper classifies the discharge waveforms into more types, each type has a different impact on the process output parameter such as material removal, electrode wear, pollution of the gap, stability, etc. We used six voltage features to derive the pulse generator matrix. According to this matrix, we can write eighteen distinguished Boolean expressions. Any pulse type can be expressed by a Boolean representation of required pulse fearture bits. After we identified the pulse types, this paper also proposes an algorithm to predict the next pulse state, and then construct the fuzzy control rule table for the use of EDM adaptive control.