摘要
该文针对直流互联异步电网的频率稳定问题,考虑数据驱动方法在电力系统紧急控制中的时效性优势,提出一种基于多层支持向量机(support vector machines,SVM)的交直流电网频率稳定紧急控制方法。该方法利用3层SVM模型实现直流紧急功率支援和自动切负荷控制相结合的最优频率稳定协调控制策略。其中,第1层是建立基于v-SVR的扰动后频率稳定预测模型,该模型能够基于扰动后瞬间的WAMS数据快速预测出稳态频率;第2层是建立基于C-SVC的频率稳定控制方式判断模型,该模型根据预测频率判断相应采取的合理控制方式;第3层是建立基于v-SVR的频率稳定控制策略制定模型,该模型根据所选择的控制方式制定出最优控制策略。仿真分析表明,该方法不仅大大提高了控制时效性,而且具有很好的准确性和有效性,适用于交直流电网频率稳定紧急控制的在线应用。
In this paper, aiming at the frequency stability problem of the asynchronous power grid with DC-interconnected, considering the time-effectiveness advantage of the data-driven methods in emergency control, a frequency stability emergency control method based on multi-layer support vector machines (SVM) was proposed. This method mainly implements the frequency stability control coordinated with emergency DC power support and automatic load-shedding according to a three-layer SVM model. The first layer is to establish a prediction model based on v-SVR, which can quickly predict the post-disturbance steady-state frequency according to the WAMS data. The second layer is to establish the judgment model of frequency control mode based on C-SVC, which can judge the reasonable control mode based on the predicted frequency. The last layer is to establish a frequency stability control scheme formulation model based on v-SVR, which can formulate an optimal control scheme for the judged control mode. Simulation analysis shows that the proposed method not only greatly improves the timeliness of the control, but also has good accuracy and effectiveness. It is suitable for on-line frequency stability emergency control.
作者
胡益
王晓茹
滕予非
艾鹏
车玉龙
HU Yi;WANG Xiaoru;TENG Yufei;AI Peng;CHE Yulong(School of Electrical Engineering, Southwest Jiaotong University, Chengdu 610031, Sichuan Province, China;State Grid Sichuan Electrical Power Research Institute, Chengdu 610072, Sichuan Province, China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2019年第14期4104-4117,共14页
Proceedings of the CSEE
基金
中国博士后科学基金资助项目(2015M582543)
四川省重点研发计划项目(2017GZ0054)~~
关键词
直流互联异步电网
支持向量机
频率预测
直流紧急功率支援
切负荷控制
频率稳定控制
asynchronous power grid with DC-interconnected
support vector machines (SVM)
frequency prediction
emergency DC power support
load shedding
frequency emergency control