Reliable Maintenance Scheduling of Pumps in Existing Thermal Power Stations

Abstract A reliable decision support system for determining maintenance scheduling of pumps in existing thermal power stations was proposed in order to reduce the maintenance cost by prolonging the present availability. The problems due to unavailability of enough data under many different operational environments of power system pumps were solved using the dimensional reduction method with suitable modifications. Weibull distribution was used to extrapolate the future failures from the presently available data which only represented a short time window with respect to their total life time. The analysis finally determined the optimal maintenance interval of each individual pump types in the pump system by considering the system availability in order to reduce the cost of unnecessarily scheduled maintenance. The method of determining maintenance intervals of this paper can be applied to any existing systems in thermal power stations with the features of few data and many parameters, and proved to be an assistant tool for maintenance engineers in thermal power stations.