Fault Diagnosis Of Power Transformer Based On Extreme Learning Machine Optimized By Improved Grey Wolf Optimization Algorithm

For power transformers, the gas content in oil is used as the fault input feature quantity, and the accuracy of diagnosis results is not satisfactory. The problem of low accuracy of optimized extreme learning machine (ELM) of grey wolf optimization (GWO) algorithm is proposed, and a hybrid intelligent fault diagnosis method based on random forest and improved optimized extreme learning machine of grey wolf optimization algorithm is proposed. Firstly, the importance of the candidate gas ratios is score by random forest and reassembled into five groups of feature parameters in order of importance from highest to lowest and used as the input feature quantity of the model. Secondly, the extreme learning machine is optimized to randomly generate weights and thresholds using the improved grey wolf optimization algorithm to improve the prediction accuracy of the model. Finally, the simulation experiments and comparative test analysis show that the fault diagnosis model has particular effectiveness in transformer fault diagnosis.