The main challenge in maintenance planning lies in the realistic modeling of the maintenance policy. This paper is focused on the maintenance optimization of complex repairable systems using Bayesian networks. A new policy is developed for periodic imperfect preventive maintenance policy with minimal repair at failure; this policy allows us to take into consideration several types of preventive maintenance with different efficiency levels. The Bayesian networks are used for complex system modeling, allowing the evaluation of the model parameters. The Weibull parameters and the maintenance efficiency are evaluated thanks to the proposed methodology using Bayesian inference. The approach developed in this paper is applied on a real system, to determine the optimal maintenance plan for a turbo-pump in oil industry. in Wiley Online Library
Model formulationWe note by τ the operating time (i.e. up-time) over the time interval T. Each PR type reduces the failure rate of the system by an amount, which is the same for each PR type, but different from one PR type to the other. Let the initial failure rate function of the system be noted λ 0 (τ), the PM policy is described as follows ( Figure 1):The proposed optimization approach, represented in Figure 4, is based on the main following steps:1. The studied system is modeled as a BN by using the corresponding FT, the expert judgment, or both. In this BN, all the system components are represented by root nodes and connected to the other nodes, in order to represent and quantify the dependencies between the system components. The other events, which can affect the system such as the environment or functioning conditions, can be added to the BN. 2. The prior probability tables of all the root nodes are generated using Equation (11) or expert judgment if the corresponding node does not represent a system component. The CPT are generated from the FT gate and the expert judgment. 3. The CDF of the whole system is obtained from Bayesian inference, and its corresponding parameters are assessed by using linear regression (Equation (12)).
7574. The PM efficiency factors are computed for each PR type, from Equation (16). 5. The optimal time interval between two successive PM actions (i.e. τ *) is obtained by minimizing the expected maintenance cost per unit time (Equation (9)). This step is performed for several values of K. For large systems, the use of combinatory optimization algorithms can be recommended. 6. The optimal PM plan corresponds to the couple (τ *, K), which leads to the lowest maintenance cost.One can see that the use of BNs in the optimization approach allows the combination of historical data with expert judgment for the estimation of reliability and maintenance efficiency parameters. In addition, the approach can be applied for complex systems with the consideration of several PM types with specific effect for each PM type on one hand, and the consideration of the system components on the other hand.
Industrial applicationIn order to validate the proposed model, the...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.