基于小波神经网络的双电极同步伺服放电加工工艺效果预测  

Prediction for electrical discharge machining process with synchronous servo double electrodes based on wavelet neural network

作　　者：于丽丽[1] 刘永红[1] 蔡宝平[1] 朱连章[2] 纪仁杰[1] 董欣[1] 

机构地区：[1]中国石油大学机电工程学院,山东东营257061 [2]中国石油大学计算机与通信工程学院,山东东营257061

出　　处：《中国石油大学学报（自然科学版）》2008年第4期87-90,共4页Journal of China University of Petroleum(Edition of Natural Science)

基　　金：国家自然科学基金项目(50675225);山东省科技攻关项目(2006GG2204001)

摘　　要：针对非导电工程陶瓷双电极同步伺服放电加工工艺参数与加工效果间的高度非线性,提出了一种既能充分利用神经网络的自学习能力,又能利用小波良好的时频局部化特性的非导电工程陶瓷双电极同步伺服放电加工效果预测的小波神经网络方法,并建立了预测模型,同时将预测结果与传统神经网络模型的预测结果进行了比较。结果表明,小波网络模型的收敛速度和预测精度均优于传统神经网络模型。According to the high nonlinear feature and complex nature of electrical discharge machining process with synchronous servo double electrodes for non-conductive engineering ceramics, a forecast method based on wavelet neutral network which can make full use of part characteristics of wavelet time-frequent and self-study ability of neutral network was presented and the forecast model was set up. The forecast results by wavelet neural network model were compared with those of traditional neural network model. The results show that the model based on wavelet neural network is better than that based on traditional neural network inboth convergence rate and prediction accuracy.

关 键 词：小波网络 人工神经网络 非导电工程陶瓷 放电加工 双电极同步伺服 

分 类 号：TG661[金属学及工艺—金属切削加工及机床]