基于改进GSA-SVM模型的电力变压器故障诊断

准确评估输变电设备运行状态是电力企业生产技术工作的核心内容。为提高电力变压器故障诊断精度，避免传统引力搜索算法（GSA）自身收敛速度慢且易陷入局部最优区等不足，提出一种利用混沌序列改进GSA的支持向量机（SVM）模型，用于电力变压器故障诊断中。首先利用混沌序列来增加重力粒子的多样性，目的是避免在其训练时陷入局部最优区；然后利用改进的GSA算法来优化SVM模型自身的参数，从而提升该模型的预测准确率；最后将预测结果与其他3种传统诊断模型的预测结果进行了对比分析，结果表明利用混沌序列改进的GSA-SVM模型有着更好的泛化能力以及更高的分类准确率。

Power Transformer Fault Diagnosis Based on Improved GSA-SVM Model

Accurate assessment of power transmission equipment operation status is the core content of power enterprises production & technology . In order to improve the accuracy of power transformer fault diagnosis， avoid the slow convergence speed of traditional gravity search algorithm（GSA） and the tendency to fall into local optimum area，support vector machine （SVM） model is proposed， which uses chaotic sequences to improve the diversity of gravity particles for preventing to fall into local optimum area. Then improved GSA algorithm is used to optimize the parameters of SVM model， so as to improve the prediction accuracy. Finally， the prediction results are compared with those of other three traditional diagnostic models. The results show that the GSA-SVM model based on chaotic sequence has better generalization ability and higher classification accuracy.