System Reliability and Cost Optimization Under Various Scenarios Using NSGA-III

Nowadays, industrial systems need to be as reliable as possible in order to ensure safety and competitiveness. This paper addresses the reliability-redundancy allocation problem (RRAP) of an overspeed protection system in a power plant under various scenarios. Previously, this kind of optimization problems were solved using mathematical programming techniques and considered as a single objective optimization problem, however more recently, bio-inspired algorithms are used to solve this type of optimization problem. In the present work, a multi-objective evolutionary optimization algorithm, called the non-dominated sorting genetic algorithm (NSGA-III) is implemented to solve the problem under a set of nonlinear design constraints. The NSGA-III demonstrates its ability to generate a set of non-dominated solutions. The results are discussed under various scenarios of minimum allowable reliability.