基于RBF-SVR组合模型的切削力预测  

Cutting Force Forecast Based on RBF-SVM Combined Model

作　　者：于洋[1] 王关 贾智旗 任玉斌 Yu Yang;Wang Guan;Jia Zhiqi;Ren Yubin(College of Mechanical Engineering,Xi′an University of Science and Technology,Xi′an 710054,China)

机构地区：[1]西安科技大学机械工程学院

出　　处：《工具技术》2022年第8期92-96,共5页Tool Engineering

基　　金：陕西省秦创原科技创新专项(2021ZDZX-GY-0006)。

摘　　要：为准确地预测机械加工中切削力的变化,提出了一种基于RBF神经网络和支持向量机回归的RBF-SVR组合预测模型。利用RBF神经网络非线性拟合能力强、支持向量机回归处理小样本数据的特点分别预测切削力,采用最优加权法对预测值进行处理,以确定权重系数并构建组合预测模型,运用三种误差评价指标评价模型预测精度,并在相同切削力数据的前提下,与其他文献建立的AGEM组合模型预测值作对比。研究结果表明:RBF-SVR组合模型的误差指标均小于单一模型,其切削力预测值与真实值吻合效果更好;相对于AGEM组合模型,三种误差评价指标的误差值最少降低了40%左右,说明RBF-SVR组合预测模型在预测切削力时的精度更高、泛化性更好。In order to accurately predict the change of cutting force in machining,RBF-SVR combined prediction model based on RBF neural network and support vector machine regression is proposed.RBF neural network has strong nonlinear fitting ability,the support vector machine regression has strong ability to process small sample data,the prediction of cutting force is carried out by ueing them respectively.For building a combined prediction model,the optimal weighting method is used to process the predicted values to determine the weight coefficient to.Three error evaluation indexes are used to evaluate the prediction accuracy of the model.On the premise of analyzing the same cutting force data,the predicted values are compared with those of AGEM combined models established by other literatures.The results show that the error index of RBF-SVR combined model is smaller than that of single model,and the predicted value of cutting force is in better agreement with the real value.Compared with AGEM combination model,three error evaluation indexes are reduced by at least about 40%,which shows that RBF-SVR combination prediction model has higher precision and better generalization in predicting cutting force.

关 键 词：切削力预测 小样本 组合预测模型 RBF神经网络 支持向量机回归 

分 类 号：TG501.3[金属学及工艺—金属切削加工及机床]