基于改进k均值聚类算法的可控负荷调度方法研究

Research on Controllable Load Method Based on Improved K-means Clustering

中文关键词: 用户需求  k均值算法  聚类分析  负荷调度

英文关键词: Demand response  k-means  Clustering algorithm  load scheduling

中文摘要:

      本文首先对电力市场中的可控负荷按照负荷特点进行了聚类划分，将市场中的用户划分为不同类型，并设计相应的电价激励机制，从而实现用户电力负荷的调整。由于聚类算法的优劣程度很大程度上受到初始聚类中心选择的影响，因此，本文在基于k均值聚类算法的分析基础上，首先分析了不同电力负荷数据的密度特点，再通过比较，选取密度大的数据点来构建负荷的初始聚类中心。通过对聚类结果进行分析，制定对可控负荷的调度策略，实现不同负荷的不同电价激励，从而达到对可控负荷的调度的目的。该方法可以有效减小在聚类时的迭代次数，并获得稳定的聚类结果。仿真结果表明，本文提出的方法可以节约系统供电成本，有效地鼓励用户积极主动的参与对电力市场的负荷调节。

英文摘要:

      The controllable load in the power market is clustered according to the load characteristics, and the users in the market are divided into different types, and the corresponding price incentive mechanism is designed to adjust the power load. The selection of initial clustering centers will directly affect the quality of load clustering. Based on K-means algorithm, this paper proposes a method to construct initial clustering centers based on the density characteristics of load data. This method can effectively reduce the number of iterations and obtain stable clustering results. In this paper, the clustering results are used as the scheduling strategy for controllable load, and the controllable load can be dispatched by formulating different price incentives. The simulation results show that the improved clustering method can save the cost of power supply and effectively promote the active participation of users in load regulation of power market.