Flexible Generator Maintenance Scheduling in a Practical System Using Fuzzy Logic and Genetic Algorithm

Abstract

A hybrid fuzzy knowledge based system and genetic algorithm has been proposed to solve the maintenance-scheduling problem for thermal generating units. In contrast to the existing maintenance scheduling methods, the proposed approach implements fuzzy knowledge based system to emulate the power plant personnel’s experiences considering the uncertainties in the constraints, such as, the operating hours of the generators, specific fuel consumption, crew and resources availability. The genetic algorithm optimises the total generating cost and the maintenance cost as the objective functions. Simulations were carried out on a practical thermal power plant consisting of 19 generating units, over a six month planning horizon. The performance of this hybrid genetic algorithm-fuzzy knowledge based system is compared to the classical method currently in use at the thermal power system. The results obtained show that this hybrid approach is an effective and practical way in solving the maintenance scheduling problem. Moreover, the method proposed is easily revisable and able to handle larger systems and longer planning horizons in a considerable computational time.