Research of turbine generator unit fault diagnosis method based on CPN

Abstract

The theory of the counter propagation network (CPN) is introduced in this paper, and a fault diagnosis method of turbine generator unit based on CPN is proposed, taking the turbine generator unit in thermal power plant as research object. This method serves the fault spectrum feature vectors of turbine generator unit as the learning samples to train the CPN, and then makes the network can reflect the mapping relationship between the fault spectrum feature vectors and the fault types to achieve the goal of diagnosis. The simulation results show that: compared with the BP neural network, the CPN is featured by stronger ability of classification and recognition and better immunity, so that it can diagnose the vibration faults of turbine generator unit effectively, and has a certain theoretical significance and application value.