Excitation fault diagnosis method and technology based on knowledge base

The invention discloses a knowledge base-based excitation fault diagnosis method and process. The method comprises the following steps of: 1, acquiring to-be-diagnosed state data of an excitation system according to a time sequence; preprocessing the to-be-diagnosed state data of the excitation system; 2, performing diagnosis in a pre-trained LSTM-CNN hybrid neural network fault diagnosis model in a knowledge base to obtain a fault diagnosis result of the excitation system; and step 3, according to the working state information of the generator excitation system acquired by the sensor acquisition unit through an on-line monitoring processing module in the knowledge base, adopting a fault tree analysis method to carry out early warning on the fault of the generator excitation system, or carrying out diagnosis on the fault of the generator excitation system. According to the invention, the fault of the excitation system can be diagnosed in various modes, so that the accuracy and convenience of fault diagnosis a