基于神经网络的超声无磨料外圆抛光加工建模与仿真  

Modeling and simulation of ultrasonic external circle polishing without abraser based on neural network

作　　者：苏玉珍[1] 聂孟辉 

机构地区：[1]郑州大学机械工程学院,河南郑州450001 [2]洛阳钼业集团股份有限公司

出　　处：《矿山机械》2009年第10期23-26,共4页Mining & Processing Equipment

基　　金：国家自然科学基金资助项目(10176010)

摘　　要：为了精确预测超声无磨料外圆抛光加工的效果,应用神经网络理论建立了超声无磨料外圆抛光加工件表面粗糙度的RBF神经网络模型。根据实际情况,进行了网络结构设计,比较了不同特征参数时网络模型的性能,并对所建模型进行了仿真验证,结果表明,预测数据与实测数据有较好的一致性,而且RBF网络性能稳定,说明可以使用神经网络模型对超声无磨料外圆抛光加工效果进行预测。In order to correctly forecast the effects, the RBF neural network model is established for the surface roughness after external circle polishing without abraser ultrasonic machining. The structure of network is designed according to the practical conditions, and the performance of the model is compared under different feature parameters. The model is validated by simulation and the results show that the data of forecast are in accord with that of test, and the RBF has a stable performance, which provides that the effect of ultrasonic external circle polishing without abraser can be forecasted by the neural network model.

关 键 词：神经网络 超声加工 抛光 仿真 

分 类 号：TG663[金属学及工艺—金属切削加工及机床]