Application of Synthetic Neural Network for Fault Diagnosis of Steam Turbine Flow Passage

It is difficult to determine the normal value for fault diagnosis of steam turbine flow passage, because the flow passage fault or regenerative system fault may result in the decrease of the relative internal efficiency of steam turbine. Based on analyzing the feature of flow passage and regenerative system, the flow passage condition of steam turbine can be evaluated by relative internal efficiency of stage groups. According to various effect factors, the methods to determine the normal value of relative internal efficiency of stage groups is proposed in this paper, which include application of synthetic neural network. Compared the measured values with these normal values, the operating condition of steam turbine flow passage can be evaluated and the detailed reasons of fault can be diagnosed.