A new method of turbine-generator vibration fault diagnosis based on correlation dimension and ANN

摘要

This paper analyzes generator stator vibration frequency spectrum characteristic and correlation dimension characteristic based on phase space reconstruction of fractal theory, when generator is in three conditions of normal operation, rotor excitation winding short circuit and stator winding fault. The results show that rotor excitation winding short circuit and stator winding fault resemble stator vibration frequency spectrum, but three conditions have a different correlation dimension. So a new method of generator vibration fault diagnosis based on correlation dimension and artificial neural network (ANN) is proposed, which selects the stator vibration correlation dimension for the input vector of ANN, and practically acquired MJF-30-6 generator data for learning samples. The results of verification show that the method can identify efficiently three generator conditions.