基于机器学习的分布式发电并网保护研究  

Research on Distributed Generation Grid Connection Protection Based on Machine Learning

作　　者：白明辉[1] 袁绍军[1] BAI Minghui;YUAN Shaojun(Chengde Power Supply Company of State Grid Jibei Power Co.,Ltd.,Chengde 067000,China)

机构地区：[1]国网冀北电力有限公司承德供电公司,河北承德067000

出　　处：《微型电脑应用》2023年第12期73-76,共4页Microcomputer Applications

基　　金：国网河北供电公司2020承载力调度三类分布式调度管理系统正常项目(GNRX202100643)。

摘　　要：由于分布式发电的接入和退出随时可能发生,对电网的影响也更为复杂,因此为了确保分布式发电并网系统的稳定运行,提出基于机器学习的分布式发电并网保护方法。针对分布式发电并网系统中存在故障、扰动以及孤岛事件,利用机器学习算法中的支持向量机建立分类器模型,对上述事件进行分类检测;设计双重滤波的协调控制策略,提高分布式发电并网在扰动事件和故障事件下的功率稳定性;通过变系数下垂控制策略,对系统直流母线电压控制,提高系统在孤岛运行状态下的电流质量。实验结果表明,所提方法可有效控制分布式发电并网的功率,保持电流稳定,具有良好的事件识别能力和保护能力。As the access and exit of distributed generation occur at any time,the impact on the power grid is more complex.In order to ensure the stable operation of distributed generation grid connected system,a method of distributed generation grid connected protection based on machine learning is proposed.In view of the fault,disturbance and islanding events in the grid connected distributed generation system,the support vector machine in the machine learning algorithm is used to establish a classifier model to classify and detect the above events.The coordinated control strategy of dual filtering is designed to improve the power stability of distributed generation grid connection under disturbance events and fault events.By a variable coefficient droop control strategy,the DC bus voltage of the system is controlled to improve the current quality of the system during islanding operation.The experimental results show that the proposed method can effectively control the power of distributed generation grid connection and maintain the current stability,have good event identification ability and protection ability.

关 键 词：机器学习 支持向量机 双重滤波 分布式发电并网 下垂控制 

分 类 号：TM315[电气工程—电机]