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 .
出处
《机床与液压》
北大核心
2006年第10期52-54,共3页
Machine Tool & Hydraulics
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
Genetic algorithm
Neural networks
Engineering ceramics
WEDM
Parameters optimizing
