Fault diagnosis of rotor winding inter-turn short circuit in turbine-generator based on BP neural network
Yonggang Li, Yanjun Zhao, Lei Chen, Xuan Ji · International Conference on Electrical Machines and Systems · 2008
The electromagnetic characteristic and rotor vibration characteristic of turbine-generator are analyzed when rotor winding inter-turn short circuit fault has happened. This paper reveals that exciting magnetic force Ff is constant in a fixed condition whereas the exciting current If increases in case of rotor inter-turn fault. This paper also finds relevant characteristic parameters. Based on the theory, we can get training patterns without doing destructive tests. Then BP (back propagation) neural network can be adequately trained and diagnosis rotor winding inter-turn short circuit. BP neural network is independent on mathematic models and parameters of turbine-generator. Finally practically acquired dynamic experiment data of the MJF-30-6 generator, the results of verification show that the theory analysis is right and the method is efficient and accurate.