Transformer fault diagnosis based on improved dung beetle algorithm optimizing SVM

When using support vector machine (SVM) for transformer fault diagnosis, it is difficult to choose the optimal values of the kernel parameters g and c based on manual experience, resulting in lower accuracy of fault diagnosis. The Dung Beetle Optimizer (DBO) algorithm has the drawbacks of getting trapped in local optima and low convergence accuracy in later stages. To address these two issues, an improved Dung Beetle Optimizer (IDBO) is proposed to optimize the transformer diagnostic model for SVM parameters g and c. This method improves and optimizes DBO using the tent chaotic mapping, thereby enhancing the algorithm's optimization capability and accelerating convergence speed. Through comparative testing with the original DBO, Sparrow Search Algorithm (SSA), Grey Wolf Optimizer (GWO), the superiority of the IDBO algorithm is verified. Compare the IDBO-SVM fault diagnosis model with SVM, SSA-SVM, and GWO-SVM models in fault diagnosis experiments, and compare it with the DBO-SVM model in stability simulation experiments. The results indicate that IDBO-SVM model has the best accuracy, better stability, and faster diagnosis speed, resulting in better fault diagnosis performance.