摘要
变压器在电力系统中发挥着极其重要的作用,由于其工作的特殊性,容易发生过热、放电等故障。因此,如何精确检测变压器故障类型尤为重要。目前最常用的分析方法为油中溶解气体分析法(DGA)。针对DGA存在着编码缺失、故障特征少且分布不均等问题,提出一种多策略改进旅鼠算法(IALA)优化深度混合核极限学习机(DHKELM)的变压器故障诊断方法。首先,使用核主成分分析法(KPCA)对变压器故障数据集进行预处理;其次,引入Singer混沌映射、柯西分布扰动策略、循环变异策略对旅鼠算法(ALA)进行优化,提高算法的全局搜索能力;然后,将IALA应用于DHKELM的参数寻优,构建IALA-DHKELM模型;最后,与其他5种模型进行对比试验,实验结果表明该方法提出的准确率为97.4359%,优于其他诊断模型。
Transformers play a crucial role in power systems and are prone to faults such as overheating and discharge due to their unique operational characteristics.Thus,precise detection of transformer faults is essential.Currently,dissolved gas analysis(DGA)is the most widely used diagnostic technique.To address issues in DGA such as encoding deficiencies,limited fault features,and uneven data distribution,this study proposes a transformer fault diagnosis method based on a multi-strategy Improved Arctic Lemming Algorithm(IALA)optimized Deep Hybrid Kernel Extreme Learning Machine(DHKELM).Firstly,the transformer fault dataset is preprocessed by nuclear principal component analysis(KPCA).Secondly,Singer chaos mapping,Cauchy distribution perturbation strategy and cyclic mutation strategy are introduced to optimize the lemming algorithm(ALA)to improve the global search ability of the algorithm.Then,IALA is applied to DHKELM parameter optimization to construct the IALA-DHKELM model.Finally,compared with other five models,the experimental results show that the proposed method has an accuracy of 97.4359%,which is better than other diagnostic models.
作者
孙强
孙霞
李文清
Sun Qiang;Sun Xia;Li Wenqing(School of Electrical and Information Engineering,Anhui University of Science and Technology,Huainan,Anhui 232001,China)
出处
《黑龙江工业学院学报(综合版)》
2025年第10期111-117,共7页
Journal of Heilongjiang University of Technology(Comprehensive Edition)
基金
国家自然科学基金项目(项目编号:51874010)
安徽省质量工程项目(项目编号:2020xsxxkc142)。
关键词
变压器故障诊断
核主成分分析法
旅鼠优化算法
深度混合核极限学习机
transformer fault diagnosis
kernel principal component analysis
lemming optimization algorithm
deep hybrid kernel extreme learning machine
