Remaining Useful Life Prediction for Heterogeneous-Wearing Cutting Tools Based on Gamma Process and Bayesian Inference

Accurate prediction of tool remaining useful life (RUL) is essential to ensure high product quality and machine reliability. Considering that the tool wearing processes are stochastic and heterogeneous, this paper proposes a RUL prediction method for cutting tools based on the Gamma process and Bayesian inference. The tool wearing process is modeled by the Gamma process, and the average wearing rate is assumed to follow the inverse gamma distribution for characterizing the heterogeneity in wear. Then, the posterior distribution of the average wearing rate is dynamically updated using the online wear data of an operating tool based on the Bayesian inference. Further, an approximate expression for the probability density function of the tool RUL is derived based on the Poisson distribution. Finally, the proposed method is applied to a wear data set consisting of 16 milling tools. The results show that the proposed method can accurately predict the future degradation trajectory and the RUL of each individual cutting tool.