摘要
针对分时电价下电力供需不平衡,导致资源浪费及用电高峰期的电网压力过大问题,本文提出了基于分时电价下电力需求响应分析。首先,对电力需求的行为进行了分析,分时电价由峰、平、谷时段电价的不同构成,电力用户也会根据3个阶段电价的不同,结合其他影响因素,制定用电措施,并初步提出了电力需求响应的模型包含的影响因素;其次提出了改进的支持向量机回归模型(SVM),采用网格法优化SVM里面的δ和C算子,提升SVM的预测精度,将对电力需求侧的影响因素分为输入、输出因子,构建基于SVM的分时电价下电力需求响应模型;最后,针对实际企业进行算例仿真,从MATLAB仿真结果中可以看出,本文所提方法精度很高,在分时电价下电力需求侧的用电量预测中具有高可靠性和实用性,为决策人员制定合理的供需计划提供了可靠的理论基础。
In view of the time-sharing electricity power imbalance between supply and demand, lead to waste of resources and the peak season of power grid stress problem, is proposed in this paper, based on electric power demand response analysis under time-sharing electricity price. Firstly, the behavior of electricity demand is analyzed, and the different components of the electrlcity price are divided into different components of the electricity price. The power users will also formulate the electricity measures according to the different prices of the three stages, combine other factors, and prelim- inarily put forward the influencing factors of the model of power demand response. Secondly, the improved support vector machine regression model （SVM） is proposed. The inside of the grid method is used to optimize the SVM 8 and C operator, impro＇#e tt^e （o^ecast precision of SVM, the influencing factors on electric power demand side can he divided in- to the input and output factors, build the time-sharing electricity price under the electric power demand response model based on SVM. Finally, an example simulation of the actual enterprise is carried out. As can be seen from MATLAB simulation results, the accuracy of the proposed method is very high. In the forecast of power consumption of power de- mand side at the time of separation o{ electricity price, it has high reliability and practicability. It provides a reliable theo- retical basis for decision-making personnel to formulate reasonable supply and demand plan.
出处
《国外电子测量技术》
2017年第12期10-13,共4页
Foreign Electronic Measurement Technology
基金
国家电网公司科技项目(SGZJ0000BGJS1500460)资助
关键词
分时电价
电力需求
响应
SVM
time-sharing price
electricity demand
response
SVM
