Cox regression model applied to Pitot tube survival data
Résumé
The Pitot tube is a pressure measurement sensor used to measure flow fluid velocity. Cox proportional hazard model is widely used in analysis of time to event data. Model is applied to study the probability of failure of a component at different times. The idea of the model is to define this probability as a product between a baseline hazard function and covariates-related component. In this paper, we develop proportional hazard models to estimate Pitot tube reliability and hazard rate in a finite time horizon. The reliability estimated function is used after to define the optimal number of preventive maintenance actions in a finite time horizon. Two groups of models are considered: first group ignores interactions between covariates and the second takes these interactions into account.