الخلاصة:
Our thesis is concerned with the development of realistic models allowing to describe the real degradation process of complex repairable systems when the maintenance and reliability data of the system are available. This description allows a better knowledge of the future performance of the system and can be used as a decision support tool for the optimization of the preventive maintenance policy. The reliability is estimated in this thesis by the generalized proportional intensity model (GPIM) which presented, among a set of imperfect repair models, the best quality of fit validated by the maximum likelihood approach. GPIM makes it possible to incorporate into the failure intensity function the effect of corrective (MC) and preventive (MP) maintenance
actions as well as the effect of various information concomitant to the systems. These are called
""covariates"" and provide more insight into the conditions under which the system operates. The results of GPIM make it possible to optimize the maintenance plan of the system taking into account its future performance influenced by all the explanatory factors. The recommended period must present, over a fixed or variable horizon, the optimum cost and availability. Programming results are obtained by simulation algorithms using the MATLAB simulation language. This thesis is part of an improvement in the operating safety of petroleum equipment, naturally closely dependent on the efficiency of maintenance operations.
