Contributions au développement de politiques de remplacement préventif pour des sytèmes multi-composants
|Abstract:||The aim of this thesis is to develop preventive replacement policies for multi-component systems. Systems are composed of several components connected under a known configuration and subject to random failures. Each replacement policy defines the actions to be taken according to the state of the system or its components and it is intended to delay the occurrence of failures and extend the lifetime of the system. From the theoretical point of view, the extension of replacement models from single-component systems to multi-component systems is not obvious. The difficulty is due primarily to the interaction or dependence between the different components of the system. In this thesis the focus has been put on the stochastic and economic dependencies between components. For stochastic dependence the propagation of the failure is modeled by the domino effect for a two-component parallel system, and two age replacement policies are investigated. In the first policy, we assumed that the cost structure is constant whereas in the second policy a variable cost structure is assumed. We proposed also a bi-objective replacement model that optimizes both expected replacement cost rate and system availability. For economic dependence, we proposed a failure counting replacement policy for a parallel system and we integrated it in a redundancy allocation model for a serie-parallel system. The mathematical model has been built taking account of this policy and Simulated Annealing algorithm has been used as resolution approach.|
|Document Type:||Thèse de doctorat|
|Open Access Date:||24 April 2018|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.