基于改进天鹰算法优化核极限学习机的油浸式变压器故障诊断研究

Fault Diagnosis Study of Oil-Immersed Transformers of Optimized Kernel Based Extreme Learning Machine Based on Modified Aquila Optimizer

为提高油浸式电力变压器故障诊断的精度及可靠性,提出了一种基于改进天鹰算法(modified aquila optimizer, MAO)优化核极限学习机(kernel based extreme learning machine, KELM)的油浸式电力变压器故障诊断方法。利用Tent混沌映射、卡方概率密度函数,对原始天鹰优化算法(Aquila Optimizer, AO)进行改进,改进后的算法有效提升了收敛速度与寻优精度。利用MAO算法对核极限学习机模型中的正则化系数和核函数参数进行联合寻优,构建最优故障诊断模型。实验结果显示,MAO-KELM对变压器故障诊断的准确率达到95.8%,比AO、GWO和PSO优化的核极限学习机故障诊断模型分别提升了3.52%、10.07%和11.64%,体现了MAO算法的优越性,同时与传统模型进行比较,证明所提方法的诊断效果具有明显优势。

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.