Stochastic Tool Wear Prediction for Sustainable Manufacturing

To provide scientific support for decision-making in critical applications such as maintenance scheduling and inventory management, tool wear monitoring and service life prediction are of significance to achieving sustainable manufacturing. Past research typically assumed time-invariant machining settings in modeling wear progression, hence is limited in accurately tracking varying wear rates. This paper presents a stochastic joint-state-and-parameter model with machining setting as a parameter that affects the state evolution or tool wear propagation. The model is embedded in a particle filter for recursive wear state prediction. Effectiveness of this method is verified through experimental data measured on a CNC milling machine.