摘要
需求响应(DR)作为电力需求侧管理(DRM)的重要措施,对提升电网安全、优化电力资源配置具有重要意义,目前已得到深化应用。为明确各类用户负荷特性和评估其参与DR的潜力,需要对电力用户进行分类。首先,综合考虑用户属性、用电时间规律等因素,建立基于集合经验模态分解(EEMD)和模糊C均值聚类(FCM)的需求响应用户负荷曲线分类模型。利用EEMD将某电网82组负荷数据分别分解为本征模态分量和趋势分量;然后,采用FCM对平稳的本征模态分量进行聚类。结果表明:EEMD-FCM模型在迭代12次后目标函数值稳定收敛于69.87,与传统FCM聚类相比,收敛值更小,求解时间更快,此外,EEMD-FCM在用户分类上比传统FCM更加精确;最后,基于用户用电负荷曲线分类结果提出不同的需求响应策略,为电网实施精准需求响应提供理论支撑。
Demand response,as an important measure of power demand-side management,is of great significance for improving grid security and optimizing the allocation of power resources,and has gradually been applied so far.In order to clarify the load characteristics of various types of users and accurately assess their potential to participate in demand response,it is necessary to classify the power users.Firstly,this paper comprehensively considers user attributes,electricity time rules and other factors,and proposes a demand response of user load curve classification model based on EEMD and FCM.EEMD was used to decompose 82 sets of load data of a certain power grid into eigenmode components and trend components,and then the FCM algorithm was used to cluster the most stable eigenmode components.The results show that the objective function value of the EEMD-FCM model stably converges to 69.87 after 12 iterations.Compared with the traditional FCM clustering,the convergence value is smaller and the solution time is faster.In addition,EEMD-FCM is more precise than traditional FCM in user classification.Finally,based on the user classification results of the power consumption characteristics,different demand response strategies are proposed to provide theoretical support for the grid to implement precise demand response.
作者
董军
陈正鹏
窦熙皓
王媛媛
聂仕麟
DONG Jun;CHEN Zhengpeng;DOU Xihao;WANG Yuanyuan;NIE Shilin(School of Economics and Management,North China Electric Power University,Beijing 102206,China)
出处
《电力科学与工程》
2021年第10期45-54,共10页
Electric Power Science and Engineering
