Prediction of the fuel impact associated with performance degradation in power plants

Diagnosis procedures are used in power plants in order to detect the presence of anomalies that affect the performance of a system and to locate the components where the anomalies have occurred. Any decision concerning maintenance strategies is conditioned by their economic convenience. The expected fuel saving that can be obtained by removing the anomalies represents an important information for such a decision. This evaluation, called prognosis, is not trivial when several anomalies occur in the plant, mainly because the anomalies change the operating conditions of all the components.In this paper, a prognosis procedure based on a productive representation of the system is proposed and applied to a system constituted by a gas turbine and a heat recovery system. Two anomalies are simulated in the turbine and in the heat recovery steam generator. The procedure allows one to predict the additional fuel consumption caused by each anomaly, with an error of less than 5%.