| With the development of CNC EDM technology, the research of ED LAYERED MILLING gradually becomes an active subject and it is paid more attentions in the field of EDM at home and abroad. Compared to traditional EDM, ED LAYERED MILLING has some obvious advantages of saving a large number of shaped electrodes, shortening production cycle, reducing processing costs, increasing the flexibility of processing and so on. Because ED MILLING belongs to EDM, it also has the problem of electrode wear which can affect the final machining accuracy. Accurate prediction and the compensation of electrode wear is a efficient method of increasing the processing accuracy.Based on the principle of equivalent wear and orthogonal experiment, a lot of experiments have been done to measure peak current, pulse duration, pulse separation, oil pressure, electrode cross section and the machining depth. The distribution effects of these parameters were analyzed to research the rules of the electrode wear of ED LAYERED MILLING. An optimization program of distribution of the processing parameters has been proposed. The impactions of various factors on electrode wear have been study and the reasons of them have been analyzed in the paper.Because ED MILLING has complex rules, many factors impact the electrode wear and the gap state always changes in the processing. It is difficult to establish a mathematical model of electrode wear. According to the nonlinear characteristics with time varying of ED MILLING, an electrode wear prediction model using RBF network was established in this paper. The network was trained using experimental data and to predict the electrode wear under certain conditions, which make it is possible to calculate the wear volume in a real-time processing and lay the foundation of dynamic compensation of the electrode wear on-line. Compared with the BP network, the result of electrode wear prediction of RBF network is better. The prediction error of RBF network was less than 8%.
|
PDF Full Text Request