关键词: 变压器, 核极限学习机, 天鹰优化算法, 故障诊断

Abstract: In order to improve the accuracy and reliability of oil-immersed power transformer fault diagnosis, an oil-immersed power transformer fault diagnosis method of optimized kernel based extreme learning machine (KELM)based on the modified aquila optimizer (MAO)is proposed in this paper. The original aquila optimizer (AO) is improved by using Tent chaotic mapping and cardinality probability density function, and the improved algorithm effectively improves the convergence speed and the accuracy of the optimization search.The optimal fault diagnosis model is constructed by jointly seeking the regularization coefficients and kernel function parameters in the KELM using the MAO algorithm. The experimental results show that the accuracy of MAO-KELM for transformer fault diagnosis reaches 95.8%, which is 3.52%, 10.07% and 11.64% higher than the optimized KELM fault diagnosis models of AO, GWO and PSO respectively, reflecting the superiority of the MAO algorithm,and comparing with the traditional model, the proposed method has obvious advantages in its diagnosis effect.

Key words: transformer, kernel based extreme learning machine(KELM), aquila optimizer(AO), fault diagnosis