Study on BPNN-GRNN-GPR Model for Tool Wear Value and Tool Life Prediction

In the field of tool wear prediction, small sample conditions can lead to low prediction accuracy, and a single model may not be sensitive to a specific tool wear stage. Therefore, a new sub-model coefficient allocation model using Artificial Bee Colony (ABC) algorithm is established. The model is composed of Back Propagation Neural Network (BPNN), Generalized Regression Neural Network (GRNN) and Gaussian Process Regression (GPR). The open data set of PHM 2010 is taken as an example, kernel principal component analysis method is used to achieve feature fusion after feature extraction. The test results indicate that the BPNN-GRNN-GPR model achieves the lowest error, which outperforms individual sub-models in terms of accuracy. The coefficient of each sub-model is determined according to its performance in the actual task, so the method proposed effectively reduces the problem of small sample conditions and the insensitivity of a single model. The findings hold significant importance for the advancement of intelligent manufacturing and the enhancement of tool wear prediction capabilities.