Oil and Gas Pipeline Fault Diagnosis Algorithm for new Energy Power Supply Based on Transfer Learning and Improved Residual Network

In order to solve the problem of low accuracy in predicting oil and gas pipeline faults under different working conditions, this paper proposes a fault diagnosis algorithm that combines transfer learning and residual network. This method transfers the model from the source working condition to the target working condition through transfer learning, realizes efficient prediction of faults under different working conditions, and combines batch normalization and regularization techniques to improve the generalization ability of the algorithm. Experimental results show that the accuracy of the algorithm proposed in this paper on the source working condition data training set, target working condition data training set and target working condition data validation set is $99.67 \%, {9 7. 1 1 \%}$ and 95.29% respectively, and it has the best accuracy and generalization ability under different working conditions.