基于机器学习的电力系统继电保护策略优化

打开文本图片集

中图分类号：TM77 文献标志码：A 文章编号：2095-2945（2025）26-0155-04

Abstract：Inordertoimprovetheaccuracyandspeedofrelayprotectioninpowersystems，machine learning methodssuch asneuralnetworks，supprtvector machinesandrandom forestsareusedtostudyfault identficationandprotection‘strategy optimization.Troughdatacolectionpreprocessngandmulti-modelfusionanintellintprotectionshemeisdesignedandits performanceisveifiedonasimulationplatform.Theresultsshowthattheoptimizedprotectionstrategyissignificantlybetter thanthetraditionalmethodsinfaultidentficationaccuracyandactiontime，providinganewsolutionforthesafeandstable operation of the power system.

Keywords：power system； relay protection; machine learning;fault identification; strategy

随着电力系统规模的不断扩大和复杂性的日益增加，传统继电保护策略在应对现代电力系统的动态性、非线性和不确定性方面逐渐显现出局限性。（剩余5150字）