基于GA与RBF神经网络的工程陶瓷点磨削表面硬度数值模拟  

Numerical Simulation of Surface Hardness Based on GA and RBF Neural Network in Point Grinding Engineering Ceramic

作　　者：马廉洁[1,2] 曹小兵[1] 陈小辉[1] 李琛[1] 

机构地区：[1]东北大学秦皇岛分校控制工程学院,河北秦皇岛066004 [2]东北大学机械工程与自动化学院,沈阳110819

出　　处：《组合机床与自动化加工技术》2015年第1期30-33,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金项目资助(51275083)

摘　　要：基于RBF神经网络理论,对工程陶瓷点磨削表面硬度HV与切削速度vc、进给速度f、切削深度ap、倾斜角α、和偏转角β五个工艺参数的关系进行了单因素数值拟合,并以拟合优度对拟合结果进行了检验,检验结果表明模型具有较高可信度。基于遗传优化算法,对点磨削表面硬度关于五个工艺参数的多元模型进行了优化建模,设计了正交实验对模型进行检验,最大误差在10%以内,表明模型具有较高的可靠性。Inpoint grinding engineering ceramic, the relationship between surface hardness HV and cutting speed vc, feed speedf, cutting depth ap,inclining angleαanddeflecting angleβwere univariate numerical fitting based on RBF neural network. Andfitting resultsare tested by the coefficient of determination, test results show that the model has high credibility. Amultivariate modelissolvedbetween surface hardness with fiveprocess pa-rametersbased on genetic optimization algorithmin point grinding. The orthogonal experiment was designed to test the model,the maximum error is less than 10%. The results indicated that the model has high reliability.

关 键 词：表面硬度 数值模拟 遗传算法 RBF神经网络 点磨削 工程陶瓷 

分 类 号：TH161[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]