基于KPCA-ITSO-ELM-Adaboost的变压器故障诊断方法  

Transformer fault diagnosis method based on KPCA-ITSO-ELM-Adaboost

作　　者：傅晓锦[1] 杨成 Fu Xiaojin;Yang Cheng(School of Mechanical Engineering,Shanghai Dianji University,Shanghai 201306,China)

机构地区：[1]上海电机学院机械学院,上海201306

出　　处：《国外电子测量技术》2022年第11期160-169,共10页Foreign Electronic Measurement Technology

基　　金：上海市自然科学基金(11ZR1413800)项目资助。

摘　　要：针对传统变压器故障诊断方法诊断精度低,单一智能诊断方法在实际模型中不能准确分类的问题,建立了一种改进金枪鱼算法(ITSO)优化加权极限学习机(ELM)的变压器故障诊断模型。首先,使用核主成分分析算法(KPCA)对变压器故障数据进行降维处理,去除数据中的无用信息,提高模型的识别效率,然后,利用ITSO算法对ELM进行优化,建立ITSO-ELM变压器故障诊断模型,最后,使用Adaboost算法对ITSO-ELM模型进行增强。仿真实例表明,所提方法相比于与ELM-Adaboost、TSO-ELM、ITSO-ELM模型分别提高了11.6%、7.2%、4%,验证了所提模型的有效性。Aiming at the problem that the traditional transformer fault diagnosis method has low diagnosis accuracy and a single intelligent diagnosis method can not accurately classify in the actual model, a KPCA-ITSO-ELM-AdaBoost transformer fault diagnosis model is established. Firstly, the kernel principal component analysis algorithm is used to reduce the dimension of transformer fault data, remove useless information in the data, and improve the recognition efficiency of the model. Then, the improved tuna swarm algorithm is used to optimize the limit learning machine and establish the ITSO-ELM transformer fault diagnosis model. Finally, AdaBoost algorithm is used to enhance ITSO-ELM model.The simulation results show that the proposed method improves by 11.6%, 7.2% and 4% respectively compared with ELM-AdaBoost, TSO-ELM and ITSO-ELM models, which verifies the effectiveness of the proposed model.

关 键 词：核主成分分析 金枪鱼算法 极限学习机 自适应增强算法 变压器 故障诊断 

分 类 号：TM41[电气工程—电器]