Preventive Maintenance Policy Optimization of a Nuclear Reactor High Pressure Injection System Using a Reliability-Cost Model

Previous works have demonstrated that Genetic Algorithms (GA) and Probabilistic Safety Analysis (PSA) can be applied to maintenance policy optimization problems. Here, the flexible scheduling approach proposed in former works has been enhanced. Now, operational costs are considered in the objective function, resulting in a Reliability-Cost model for evaluating industrial systems performance. The GA searches for optimum preventive maintenance policies, which considers factors, such as: probability of needing a repair, costs of related to the repair, outage times, etc. A typical Pressurized Water Reactor High Pressure Injection System has been chosen as a case study. Results ratify the good performance of the method, allowing specific analysis on the weighting factors related to costs and reliability.