Rotor fault diagnosis method based on weighted BP-AdaBoost

The invention provides a rotor fault diagnosis method based on weighted BP-AdaBoost, and relates to the field of rotor fault diagnosis. A rotor vibration signal is measured by using a sensor, rotor state features are extracted from different angles to obtain a rotor system state feature vector, then the variance of an initial feature vector is calculated, the feature vector dimension is reasonably selected according to the variance value, a BP neural network is used as a base classifier of an AdaBoost algorithm, a binary classifier is constructed, and a rotor system state feature vector is obtained. And a multi-classification classifier is constructed by combining a plurality of binary classifiers. In order to solve the problems existing during decision making of the combined multi-classification classifier, the invention provides a non-fuzzy solution coefficient based on a verification sample, K fold cross validation is combined, a weighted BP-AdaBoost multi-classifier is obtained, and rotor fault diagnosis