基于神经网络的平面磨削表面粗糙度预测模型  

Prediction Model of Surface Roughness of Plane Grinding Base on Artificial Neural Network

作　　者：杜金萍[1] 高书振[1] 王凯[1] 

机构地区：[1]河北工程大学,河北邯郸056038

出　　处：《煤矿机械》2011年第10期143-145,共3页Coal Mine Machinery

基　　金：邯郸市科学技术研究与发展计划资助项目(072112035)

摘　　要：针对平面磨削的特点,采用正交试验方法获取学习样本,用BP神经网络建立砂轮径向切入进给量、轴向进给量和工作台进给速度与表面粗糙度关系模型,并用MATLAB实现对该模型的训练和仿真,由此得出表面粗糙度预测模型。结果显示:该模型具有较高的预测精度,在学习样本的采样区间平均预测误差为3.7%,最大预测误差为7.9%。为平面磨削表面粗糙度预测提供了一种新的可行方法。Getting the learning samples by using orthogonal test method for surface grinding,and making use of BP artificial neural network model found out the relation between radial feeds,axial feeds,workbench speed and surface roughness.Through the training and simulating with MATLAB,founded the prediction model of surface roughness.The result shows that the model can get a higher accuracy of prediction.The average prediction error is 3.7%,the maximum is 7.9% in learning samples area.It provides a new-type feasible method for prediction of surface roughness.

关 键 词：平面磨削 正交试验 磨削用量 BP神经网络 表面粗糙度 

分 类 号：TH161[机械工程—机械制造及自动化] TP183[自动化与计算机技术—控制理论与控制工程]