Power Plant Maintenance Scheduling Using Ant Colony Optimization

Abstract

Under the pressure of rapid development around the globe, power demand has drastically increased during the past decade. To meet this demand, the development of power system technology has become increasingly important in order to maintain a reliable and economic electric power supply (Lin et al., 1992). One major concern of such development is the optimization of power plant maintenance scheduling. Maintenance is aimed at extending the lifetime of power generating facilities, or at least extending the mean time to the next failure for which repair costs may be significant. In addition, an effective maintenance policy can reduce the frequency of service interruptions and the consequences of these interruptions (Endrenyi et al., 2001). In other words, having an effective maintenance schedule is very important for a power system to operate economically and with high reliability. Determination of an optimum maintenance schedule is not an easy process. The difficulty lies in the high degree of interaction between several subsystems, such as commitment of generating units, economical planning and asset management. Often, an iterative negotiation is carried out between asset managers, production managers and schedule planners until a satisfactory maintenance schedule is obtained. In addition, power plant maintenance scheduling is required to be optimized with regard to a number of uncertainties, including power demand, forced outage of generating units, hydrological considerations in the case of hydropower systems and trading value forecasts in a deregulated electricity market. Consequently, the number of potential maintenance schedules is generally extremely large, requiring a systematic approach in order to ensure that optimal or near-optimal maintenance schedules are obtained within an acceptable timeframe. Over the past two decades, many studies have focused on the development of methods for optimizing maintenance schedules for power plants. Traditionally, mathematical programming approaches have been used, including dynamic programming (Yamayee et al., 1983), integer programming (Dopazo & Merrill, 1975), mixed-integer programming (Ahmad & Kothari, 2000) and the implicit enumeration algorithm (Escudero et al., 1980). More recently, metaheuristics have been favored, including genetic algorithms (GAs) (Aldridge et al., 1999), simulated annealing (SA) (Satoh & Nara, 1991) and tabu search (TS) (El-Amin et al., 2000). These methods have generally been shown to outperform mathematical programming methods and other conventional approaches in terms of the