Tool wear monitoring is of great significance to improve the machining accuracy of workpieces and the machining efficiency of production.In order to accurately predict the tool wear state,a tool wear prediction method based on multi-model weight distribution fusion is proposed.Taking the vibration signal characteristics as the re-search object,the regression tree,BP neural network and support vector regression model were used to predict the tool wear amount.By analyzing the training errors and proportions of each model,the corresponding basic probabili-ty distribution function was calculated.Using DS Evidence theory fuses the basic probability distribution functions,and finally establishes a fusion model based on the weight extraction model.By setting up comparative experiments,it is proved that the proposed method can integrate the advantages of each model,while avoiding the limitations and one-sidedness of a single model,and the decisive coefficient R2 of the experimental results is as high as 0.996 8.
关键词
刀具磨损预测/权重分配/多模型/D-S证据理论
Key words
tool wear prediction/weight distribution/multiple models/D-S evidence theory
维普
万方数据