Estimate of Cutting Tool Lifespan through Cox Proportional Hazards Model

Abstract: An important proportion of the cost of machined mechanical elements emanates from the cost of cutting tools inserts. A large part of these costs is due to improper maintenance decisions. In this paper, we demonstrate the use of a gamma process tool wear simulation for providing tool lifetimes at several cutting speeds. We show the use of a survival model (Cox Proportional Hazards model) on turning cutting inserts, using the cutting speed as covariate. Contrary to other approaches, the Proportional Hazards model is not used to produce the parameters of a reliability model, but the Mean Up Time is rather computed through the integration of the raw survival model reliability baseline. Additional simulated data is used to validate the model. The importance of a proper distribution of the PH model fitting data is highlighted by the experimental results.