基于改进聚类算法的铁芯变压电抗器故障检测

打开文本图片集

中图分类号：T391；TG323.5 文献标志码：A文章编号：1001-5922（2026）1-0292-05

Abstract：To improve the safetyof iron corevariable voltage reactors，a fault detection method for such reactors based on an improved clustering algorithm is proposed.The experimental results show that when the number of iterations reaches 6OO，the missing report rate of the improved clustering algorithm is 2.56% ，which isreduced by 90.63% ， 144.53% ，and 215.23% compared with the traditional clustering algorithm，deep learning，and Bayesian network models，respectively.The improved clustering algorithm achieves the shortest fault detection time；the detection times forwiring faults，overvoltage faults，and short-circuit faultsare35ms，31 ms，and26 ms，respectively， while the fault detection times of both deep learning and Bayesian network models exceed1OO ms.In the event of a short-circuitfault，thevibrationvelocitynear theshort-circuitedturnsincreasessignificantly.The improved clustering algorithm can efectively detect short-circuit faults in ironcore variable voltage reactors，andthe phase space reconstruction technique can further accurately detect changes in vibration velocity.The use of the grassopper optimizationalgorithm to improvethe clustering algorithm can enhance the eficiencyof randomselectionof initial clustering centers and improve the fault identification effciency of iron core variable voltage reactors.

Keywords：clustering algorithm;grasshopper optimization algorithm;iron core variablevoltage reactor;fault detection

铁芯电抗器是电力系统中最常用的电气设备之一，可以限制短路电流，抑制电容器组的谐波[1]。（剩余7293字）