基于BP神经网络的面齿轮齿面粗糙度研究  

Research of face gear surface roughness based on BP artificial neural networks

作　　者：刘金华[1] 张灵[1] 方曙光 高钦[1] 

机构地区：[1]湖南工业大学机械工程学院,湖南株洲412007

出　　处：《制造技术与机床》2017年第12期138-143,共6页Manufacturing Technology & Machine Tool

基　　金：湖南省自然科学基金项目(2015JJ5018);国家自然科学基金项目(51375161)

摘　　要：以面齿轮齿面粗糙度为研究对象,考虑到影响其磨削表面粗糙度R_a的因素很多且不明确,运用BP神经网络原理构造预测粗糙度的数学模型,通过对比模型预测值与试验值验证预测模型的精度,结果表明此模型可精确地描述砂轮转速、工件进给速度和磨削深度对面齿轮磨削加工表面粗糙度的影响。实验验证,BP神经网络模型的预测精度较高,可获得最大相对误差为9.1%的高精度预测结果。Taking the surface roughness of face gear as the study object, considered that the factors affecting the surface roughness R,, of face gear grinding are many and very. unclear, in this paper, the mathematical model of predicting roughness is constructed by using the principle of BP neural network, the accuracy of the prediction model is verified by comparing the predicted and experimental values of the model, the results show that this model can accurately describe the effect of grinding wheel speed, f^ed speed and grinding depth on the surface roughness of the grinding process. Experiment proved that the BP model can predict the surface roughness of the face gear with the maximum relative error of 9. 1%.

关 键 词：BP神经网络 面齿轮 齿面粗糙度 预测 

分 类 号：TH132.41[机械工程—机械制造及自动化]