Timing sequence feature optimized BiLSTM power transmission line fault diagnosis method

The invention discloses a time sequence feature optimized BiLSTM power transmission line fault diagnosis method. The method comprises the steps of 1, obtaining M power transmission line operation samples under n fault influence factors; 2, screening out key factors by adopting a comprehensive weight setting method to obtain a sample set under key fault influence factors, and labeling line fault labels; 3, constructing an improved BiLSTM network, wherein the improved BiLSTM network comprises a forgetting gate, an input gate, an output gate, a time sequence feature optimization module and a diagnosis module; introducing a similarity penalty term into the loss function to enhance the learning ability of feature information in different directions of the model; training the improved BiLSTM network by using a gradient descent method to obtain a power transmission line fault diagnosis model; and 4, based on the multi-dimensional monitoring data of the power transmission line acquired in real time, performing fault