Fault diagnosis system for hydraulic turbine generator

A fault diagnosis system has been developed to prevent serious accidents in a hydraulic turbine generator by detecting anomalies in their early stages. The system predicts the bearing temperature change by using the physical dynamic model of the bearing related components, such as the bearing pad and lubricating oil cooler. The unmeasurable parameter values in the model, e.g. bearing gap width, are estimated on-line by minimizing the differences between the measured and calculated values of bearing and oil temperatures using a nonlinear optimization method. The system detects and diagnoses the anomalies based on the prediction results and estimated parameter values.After evaluating the feasibility using a 1/3 scale test facility, the system was installed in the Okutataragi Pumped Storage Power Plant of the Kansai Electric Power Co., Inc. Accurate prediction performance was confirmed for the system which is now being used for commercial operation.