摘要
由于现代处理器的计算能力足以处理电网各区段各个运行模式所需参数的统计工作,因此开发识别应急模式的全新算法,并将其应用于电保护及自动化(RPA)中。本文分析了经典机器学习算法在RPA任务中的应用,特别是k-nearst、Logit模型和支持向量方法。研究了专门的可训练触发元件的使用,既用于建立新的保护,也用于提高传统类型继电保护装置的稳定性。开发的多参数RPA触发元件有助于提高事故的敏感性和识别能力,提出的应急模式识别方法适用于数字化变电站智能电子设备。
Since the computing power of modern processors is sufficient to handle the required amount of statistical data for the parameters of possible normal and emergency operating modes for each segment of the power grid,therefore,the development of a new algorithm to identify the emergency mode can be protection and automation(RPA).The application of classical machine learning algorithm in RPA task is analyzed,especially the methods of K-nearst Logit model and Support Vector Machine use of special trainable trigger elements is studied,which can be used not only to establish new protection,but also to improve the complexity of traditional type relay protection devices.The development of multi-parameter RPA trigger element is helpful to improve the accident sensitivity and recognition ability.The proposed emergency pattern recognition method is suitable for the realization of intelligent electronic equipment in digital substation.
作者
钱牧涵
QIAN Muhan(State Grid Yangzhou Power Supply Compang,Yangzhou 225000,Jiangsu,China)
出处
《电气传动自动化》
2024年第1期35-38,31,共5页
Electric Drive Automation
关键词
机器学习
继电保护及自动化
算法分析
触发元件
Machine learning
Relay protection and automation
Algorithm analysis
Trigger element
