摘要
为克服规模化风电场中磁控电抗器型SVC控制性能的局限性,避免故障切除后无功过补偿,提出了基于风电场动态电压安全决策树体系的SVC紧急控制策略。分别基于决策树中的回归树算法与分类树算法构建风机过电压预测回归树以及风机脱网分类树,形成风电场动态电压安全决策树体系。该决策树体系依据在线信息进行数据挖掘,对风机脱网状况与电压越限情况进行快速预判,并根据预测结果采取合理的电容器退出措施,避免无功过补偿。算例分析表明,所提策略不仅能为电网及风电场运行人员提供风机脱网风险信息与决策参考,还能够降低故障切除后风机因过电压而脱网的风险。
In order to overcome the control performance of magnetic control reactor (MCR)-SVC that widely used in wind farm, and avoid reactive power over-compensation after fault removal, this paper proposed a SVC emergency control strategy based on dynamic voltage security decision trees. Regression tree algorithm and classification tree algorithm have been utilized to construct over voltage prediction trees and wind turbine trip-off prediction trees respectively, which can perform data mining on online information, and make fast judgement on voltage violation and wind turbine trip-off. Furthermore, capacitor shedding is performed based on the above prediction results. Validated by simulation, the propose control strategy is able to provide supportive wind turbine off-grid information for wind farm and power system operators, and decrease the wind farm over voltage risk.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2018年第1期41-50,共10页
Proceedings of the CSEE
基金
国家自然科学基金重点项目(51537003)
黄鹤英才(科技)计划资助项目~~
关键词
决策树
风机脱网
在线预警
SVC紧急控制
decision trees
wind farm trip-off
onlinesecurity assessment
SVC emergency control