期刊文献+

基于SOM神经网络聚类的空调负荷聚合方法 被引量:14

Aggregation of Air Conditioner Load Based on Self-organizing Feature Map Neural Network
下载PDF
导出
摘要 空调类负荷的准确建模对电力系统暂态分析影响较大。为研究日益增多的空调集群特性,提出了基于自组织神经网络SOM(self-organizing feature map)聚类的空调负荷聚合建模方法。首先,通过灵敏度分析,提取对暂态分析最重要的几个空调模型参数,利用层次分析法AHP(analytic hierarchy process)确定其权重;再通过带权重训练的SOM对空调负荷进行聚类;最后,简化基于稳态模型等效变换的方法,对每一类空调进行聚合。算例表明,相比不聚类直接聚合,采用先聚类后聚合的方法对配电网中的空调负荷聚合,既可显著提高模型仿真的精度,又为研究其他负荷的聚合提供了一种新思路。 The accurate modeling of air conditioner loads has significant effect on the transient analysis of power sys- tems. To investigate the modeling and characteristics of a group of air conditioners, this paper presents a new method for aggregation of air conditioners, taking into account load clustering based on self-organizing feature map (SOM). First, several load parameters are chosen by sensitivity analysis as input vectors which are important in transient stabil- ity analysis; weight to each vector is assessed by analytic hierarchy process (AHP) ; and air conditioners are classified into different types via SOM. Last each type of air conditioners is aggregated with simplified steady state model equiva- lent method. The results indicate that the method in this paper can not only improve the simulation precision of air con- ditioners in distribution network obviously, but also contribute to the investigation of other load aggregation.
出处 《电力系统及其自动化学报》 CSCD 北大核心 2015年第11期26-33,共8页 Proceedings of the CSU-EPSA
基金 国家高科技研究发展计划(863计划)资助项目(2012AA050217) 国家电网公司重大项目(KJ[2011489])
关键词 空调聚合 自组织神经网络 层次分析法 权重训练 air conditioner aggregation self-organized neural network analytic hierarchy process(AHP) assess weight
  • 相关文献

参考文献16

二级参考文献170

共引文献388

同被引文献142

引证文献14

二级引证文献112

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部