摘要
高速电火花线切割作为特种加工技术之一,被广泛应用于各种机械加工领域。但其加工过程是一个多参数、多变化的过程,为建立一个具有一定拟合精度和泛化能力的模型,通过尝试二次通用旋转组合设计试验,以遗传算法优化BP神经网络初始权的思想建立GA-BP模型。试验表明,在相同参数设置下GA-BP比纯BP模型更具有有效性和优越性,对后续仿真系统的建立具有一定的指导作用。
The wire electrical discharge machi-ning （WEDM）is one of the nontraditional machi-ning technology,it has been widely applied in the manufacturing domain.But it is process of multipa-rameter and variable,to build a model of fitting precision and generalization ability,so the article present GA BP model based on the thought that genetic algorithm to optimize weights of BP neural network and try to use quadratic general rotation design.The experimental shows that the GA BP model is more effective and superior than single BP model under the same parameters setting,and the model has definite guidance function to further simulation system.
出处
《机械与电子》
2014年第3期41-44,共4页
Machinery & Electronics
基金
宝鸡文理学院校级重点项目(ZK11168)
