改进蜣螂优化XGBoost的变压器故障诊断研究

打开文本图片集

中图分类号：TM411 文献标志码：A 文章编号:1672-1098(2025)02-0001-09

引文格式：,,.改进蜣螂优化 XGBoost 的变压器故障诊断研究［J].安徽理工大学学报（自然科学版），2025,45(2) :1-9.

Transformer Fault Diagnosis Based on Idbo-optimized XGBoost LI Hongyue,LI Chunyu ,LI Tao

(Scholof Electricaland Information Engineering,Anhui UniversityofScienceandTechnology,Huainan Anhui 232Ool,Chna) Abstract:Objective To address the problems of difficulty in selecting key parameters for extreme gradient boosting（XGBoost）,and the dung beetle optimizer（DBO）being prone to local optima,which lead to low accuracy in transformer fault diagnosis. Methods By using 14- dimensional gas feature vectors as input,a transformer fault diagnosis model based on IDBO-optimized XGBoost was proposed.Firstly,to improve the algorithm's optimization capability,the four stages of the DBO algorithm were improved by Tent chaos mapping,adaptive spiral strategy, Cauchy-Gaussian mutation,and T-disturbance.Then,by comparative optimization tests with the original DBO, grey wolf optimizer（GWO）,sparrow search algorithm （SSA） and whale optimization algorithm（WOA）,the superiority of the IDBO algorithm was demonstrated.Finally,in fault diagnosis comparative experiments,the IDBOXGBoost model was compared with XGBoost models optimized byusing DBO,SSA and GWO,as well as with multiple machine learning methods optimized by using DBO and the IEC standard method.Results The results showed that the IDBO-XGBoost model achieved an accuracy of 91.76% and a Kappa coefficient of O. 900 8, which demonstrated better fault diagnosis effectiveness.Conclusion The IDBO-XGBoost model can effectively improve fault diagnosis accuracy,providing an effective solution for transformer fault diagnosis.

Key words : transformer;fault diagnosis;dung beetle optimizer; XGBoost

电力变压器作为电力系统中重要的一环，其稳定运行对系统正常工作具有关键性作用。（剩余14122字）