Multi-condition wear prediction and assessment of milling cutters based on linear discriminant analysis and ensemble methods

Abstract

Accurate prediction of tool wear in multi-conditions is still a thorny problem, and rapid and accurate construction of prediction models for multi-conditions is an essential part of achieving intelligent manufacturing. For multi-conditions, a novel method of rapid selection of milling tool wear prediction model based on linear discriminant analysis and the ensemble method was proposed in this paper. The linear discriminant analysis is used to divide similar working conditions, and determining the optimal prediction model for different working conditions relies on ensemble methods. Compared with the traditional transfer learning algorithm, the proposed method can be used for the rapid construction of new models and prediction of new working conditions, which are validated by real experiments. The results show that the selection of prediction models based on working condition division is superior to other models. Therefore, this method provides a guarantee for rapidly constructing accurate tool wear multi-condition predictive models.