Industrial circulating water system water supply pump cross-domain fault diagnosis method based on information fusion and meta learning

The invention provides a cross-domain fault diagnosis method for a water supply pump of an industrial circulating water system, and the method comprises the following steps: firstly, collecting multidirectional vibration signals of a driving end and a non-driving end of a water supply pump motor through a three-axis acceleration sensor, and achieving the information fusion of a plurality of sensor signals through the three-channel two-dimensional diagram of a one-dimensional vibration signal; a deep convolutional generative adversarial network is adopted to extract fault features, and an enhanced fault data set is generated; training the source domain fault data set by adopting a meta-learning method, learning common characteristics of a water supply pump fault diagnosis task, and optimizing network parameters of a target domain task; and a cross-domain fault diagnosis model is constructed, and fault diagnosis is carried out on water supply pumps of different models under complex working conditions. Aiming at