A statistical inference method for predicting the remaining useful life of milling tools based on multi-stress accelerated degradation test

Cutting tools play a pivotal role as the primary executors of mechanical processing in CNC machine tools, with their life being paramount to the overall cutting performance of the machine tools. Herein, a statistical inference approach based on a multi-stress accelerated degradation test was developed for estimating the remaining useful life (RUL) of milling tools under multi-stress working conditions. A novel class of milling tool flank wear models was first proposed, and the adaptive selection strategy of the optimal wear model was established by the criterion of minimum root mean square error (MRMSE). Subsequently, an innovative variable tool failure threshold (FT) tailored for individual milling tool was proposed, aiming at determining the pseudo life of milling tools while ensuring the quality of workpiece surfaces. The Weibull distribution model was introduced to characterize the tool life distribution. Whereafter, the relationship between stress factors and tool life information was interpreted through a series of multi-stress acceleration models, where various alternatives were provided for explicit functions representing interaction effects. Eventually, data from accelerated degradation tests of milling tool wear were utilized to validate the effectiveness of the proposed method.