Offshore wind turbine generator fault monitoring method based on multi-scale graph convolutional network

The invention discloses an offshore wind turbine generator fault monitoring method based on a multi-scale graph convolutional network. The method comprises the steps of 1, obtaining historical data of a data acquisition and monitoring control system, and performing variational mode decomposition to obtain a time domain feature matrix and an adjacent matrix on each scale; 2, establishing a time sequence diagram convolution module and a scale attention module, inputting all time domain feature matrixes and adjacent matrixes, obtaining and processing specific scale features, and finally obtaining scale attention features; and step 3, establishing a fan fault monitoring classifier aiming at realizing fault state and position monitoring, the classifier being composed of a plurality of non-linear feature mapping NFM classifiers, and inputting the attention representation of the scale into the fan fault monitoring classifier to obtain a final predicted category label. 本发明的一种基于多尺度图卷积网络的海上风电机组故障监测方法，包括：步骤1：获取数据采集与监视控制