摘要
考虑风电接入后广义负荷的不确定性,提出纵横聚类策略。针对聚类客观性与复杂场景分析需求,引入高质量的仿射传播(affinity propagation,AP)聚类算法。利用包含纵向聚类与横向聚类的纵横聚类策略对全年实测有功响应空间客观聚类,实现大尺度与小尺度数据在统一时间框架下的聚类分析:通过三步聚类实现全年大时间尺度下的纵向聚类,结果体现了细化季节特性;通过将纵向类内全部数据联排统一聚类,实现较小时间尺度下的精细横向聚类,结果体现日时段特性。引入带概率信息的广义负荷建模方法建模,并检验聚类策略的有效性。仿真结果表明,所提策略实现的聚类客观、合理,便于精确建模与现场应用,可为后续仿真分析与调度控制提供辅助参考。
Considering the uncertainty of the generalized load consisting of wind power and load, a novel longitudinal-horizontal clustering strategy was proposed. To the objectively and complicatedly analysis requirement, the high quantity affinity propagation（AP） clustering was used in this paper. Making use of the longitudinal-horizontal clustering strategy, the actual active power data were divided. The longitudinal clustering strategy can achieve the large time scale clustering and the result reflects the seasonal characteristics. The horizontal clustering strategy can achieve the small time scale clustering and the result reflects the daily time property. The generalized load modeling with probabilistic information was used to model accurately. The simulations by the actual data reveal the strategy is rational and it is conductive to modeling accurately, which is useful for dispatching and controlling considering wind power integration.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2015年第23期6028-6038,共11页
Proceedings of the CSEE
基金
国家自然科学基金资助项目(51177091)
国家863高技术基金项目(2011AA05A101)~~
关键词
仿射传播聚类算法
纵横聚类策略
广义负荷建模
细化季节特性
日时段特性
affinity propagation（AP） clustering
longitudinal-horizontal clustering strategy
generalized load modeling
seasonal characteristics
daily time property
