高压电器2023，Vol.59Issue(12)：216-222,229,8.DOI:10.13296/j.1001-1609.hva.2023.12.026
一种基于SVM算法的电力变压器机械故障智能诊断模型
Intelligent Diagnosis Model of Mechanical Fault for Power Transformer Based on SVM Algorithm
摘要
Abstract
Vibration analysis method is a non-electric diagnostic method for mechanical fault of transformer.The method can diagnose such faults as winding deformation or core looseness inside the transformer by analyzing vibra-tion characteristic of the surface of transformer oil tank.SVM,as a new machine learning method,can train and ob-tain a classification model with strong generalization ability under the condition of small samples.Based on the vibra-tion signal and SVM algorithm,an intelligent diagnosis model of transformer is constructed.The SVM training data set is constructed on the basis of collection of hundreds of groups of vibration signal data trial transformer and spec-trum feature extraction.The one classυ-SVM fault diagnosis model is trained and good results are achieved.The effec-tiveness of the model is verified,which can provide some references for the related researchers.关键词
电力变压器/振动信号/支持向量机/智能故障诊断Key words
power transformer/vibration signal/support vector machine/intelligent fault diagnosis引用本文复制引用
臧春艳,曾军,李鹏,刘宏亮,张仲程,孙路,刘耀云..一种基于SVM算法的电力变压器机械故障智能诊断模型[J].高压电器,2023,59(12):216-222,229,8.基金项目
国家电网公司2020年总部第三批科技项目(变压器热老化对绕组振动特性的影响及其抗短路能力评估策略研究).Project Supported by the 2020 Third Batch of Technology Projects of State Grid Corporation of China(Research on the Impact of Transformer Thermal Aging on Winding Vibration Characteristics and the Evaluation Strategy of Short Circuit Resistance). （变压器热老化对绕组振动特性的影响及其抗短路能力评估策略研究）
