Test Intervals, cost and risk optimization in Nuclear Reactor Safety Systems Using Cuckoo Search Metaheuristic

Technical Specifications (TSs) in the context of nuclear power plants are primarily focused on safety, ensuring that commercial nuclear reactors can safely shut down and mitigate the consequences of any postulated transient or accident. Thus, the TSs include requirements associated with safety in a systematically structured document, which describes the allowed outage time for each system component, as well as some fundamental human factors. In this context, the present work proposes a tool based on the Cuckoo Search (CS) metaheuristic to optimize the Test Intervals (TIs) of a simplified model of a High-Pressure Injection System (HPSI) of a Pressurized Water Reactor (PWR), considering the constraints and requirements established in the TSs. This study uses parameters such as test periodicity, hourly costs of testing, corrective maintenance, average test time, repair and corrective maintenance time, as well as the probabilities and failure rates of pumps and valves, in order to find an improved shutdown planning, including the simultaneous minimization of cost and unavailability. The results obtained in this study were compared with other metaheuristics from the literature, analyzing the effectiveness of each algorithm in converging to best solutions in the cost-risk trade-off. This comparative analysis shows that the developed methodology contributed not only to improving the standards of the TSs (Improved Standard Technical Specifications), but also that it can be an efficient tool for this class of problems.

Keywords: Nuclear safety, technical specifications, cuckoo search algorithm, test interval optimization, HPSI system, risk analysis..