基于GA-神经网络的工程陶瓷材料电加工参数优化研究  

Parameters Optimizing Research of WEDM Engineering Ceramics Material Based on GA-neural Networks

作　　者：骆志高[1] 徐小青[1] 王海涵[1] 

机构地区：[1]江苏大学机械工程学院,江苏镇江212013

出　　处：《机床与液压》2006年第10期52-54,共3页Machine Tool & Hydraulics

摘　　要：运用GA-高阶模糊BP神经网络对电火花线切割过程中的主要电参数(脉冲电流、脉冲宽度、工作电压、脉冲间隔、功率管数)和相应的输出参数进行学习、训练和优化,让神经网络具有预测的能力。利用优化的电参数对工程陶瓷材料A l2O3-TiC(含量30%)进行电火花线切割实验研究,得到了与GA-神经网络输出所一致的结果。A l2O3-TiC加工表面获得了最小的表面残余应力和最佳的耐磨强度。The main electrical parameters, including impulse current, impulse width, work voltage, impulse separation, the count of power tube, and homologous outputing parameters were studied, trained and optimized by applying GA - high steps faintness BP neural networks during the WEDM processing, it is showed that the neural networks have the ability of prediction. Using optimizing electrical parameters to do experiments research of WEDM engineering ceramics Al2O3 - TiC （ content 30% ）, the obtained result is accordant with the result of GA - neural networks. The least surface remnant stress and optimum wearable intension in the processing surfaces of the Al2O3 - TiC were obtained .

关 键 词：遗传算法 神经网络 工程陶瓷 电火花线切割 参数优化 

分 类 号：TG661[金属学及工艺—金属切削加工及机床] TP183[自动化与计算机技术—控制理论与控制工程]