An Intelligent Fault Diagnosis Method for Transformer Based on IPSO-gcForest

Transformers are the main equipment for power system operation. Undiagnosed faults in the internal components of the transformer will increase the downtime during operation and cause significant economic losses. Efficient and accurate transformer fault diagnosis is an important part of power grid research, which plays a key role in the safe and stable operation of the power system. Existing traditional transformer fault diagnosis methods have the problems of low accuracy, difficulty in effectively processing fault characteristic information, and superparameters that adversely affect transformer fault diagnosis. In this paper, we propose a transformer fault diagnosis method based on improved particle swarm optimization (IPSO) and multigrained cascade forest (gcForest). Considering the correlation between the characteristic gas dissolved in oil and the type of fault, firstly, the noncode ratios of the characteristic gas dissolved in the oil are determined as the characteristic parameter of the model. Then, the IPSO algorithm is used to iteratively optimize the parameters of the gcForest model and obtain the optimal parameters with the highest diagnostic accuracy. Finally, the diagnosis effect of IPSO-gcForest model under different characteristic parameters and size samples is analyzed by identification experiments and compared with that of various methods. The results show that the diagnostic effect of the model with noncode ratios as the characteristic parameter is better than DGA data, IEC ratios, and Rogers ratios. And the IPSO-gcForest model can effectively improve the accuracy of transformer fault diagnosis, thus verifying the feasibility and effectiveness of the method.