Optimization of LSSVM for Transformer Fault Diagnosis Based on KPCA and Seagull Algorithm

Regarding the issue of redundant transformer features affecting feature extraction and fault detection. This article proposes a transformer fault diagnosis method based on Kernel Principal Component Analysis (KPCA) and Seagull Algorithm optimized Least Squares Support Vector Machine (LSSVM). Firstly, the transformer fault data is preprocessed using KPCA to reduce the correlation between features and remove redundant feature components, in order to improve the accuracy of the final fault diagnosis. Secondly, in response to the problem of the influence of LSSVM parameter settings on the fault classification of the model, it is proposed to use the seagull algorithm to optimize and determine the parameters of the LSSVM model. Finally, the LSSVM model optimized by the seagull algorithm is used for final fault diagnosis, and the experimental results are compared with other models. The fault diagnosis accuracy of the proposed method in this paper is 96.33%, which is higher than several other comparison methods, verifying the effectiveness of the proposed method.