Petrochemical large unit concurrent fault diagnosis method and system

The invention discloses a petrochemical large unit concurrent fault diagnosis method, which comprises the following steps of S1, simulating various concurrent fault working conditions of a petrochemical large unit by adopting simulation software, acquiring detection data under various simulated concurrent fault working conditions, and constructing a training set; s2, collecting detection data of the petrochemical large unit under various real concurrent fault working conditions, and constructing a test set; s3, preprocessing the detection data of the test set; s4, training the convolutional neural network and the feedforward neural network by using the sample data of the training set, and performing synthetic calculation on judgment results of the two models by using a D-S evidence theory to obtain a fused fault diagnosis model; s5, inputting the preprocessed test set into the fault diagnosis model, judging the accuracy of a fault diagnosis result, and if the accuracy meets a preset value, obtaining a final f