半月刊

ISSN 1000-1026

CN 32-1180/TP

+高级检索 English
基于贝叶斯推断的用户偏好学习与响应优化
作者:
作者单位:

1.上海理工大学机械工程学院,上海市 200093;2.上海电力大学电气工程学院,上海市 200090

摘要:

灵活性负荷聚合商通过需求侧响应整合用户侧资源参与电力市场,提高系统可靠性和经济性并获得收益。但用户对用电方式和成本的偏好将对其响应电量带来不确定性,进而影响聚合商的电量申报精度和市场收益。文中将参与需求响应的负荷资源作为广义需求侧资源,构建用户响应偏好模型。进而利用贝叶斯推断对用户偏好进行学习,获得响应电量的概率性估计,生成最优响应计划。最后,建立基于偏差电量考核的聚合商市场收益模型,采用美国PJM市场用户报价及典型日交易数据进行算例仿真。仿真结果验证了贝叶斯推断对用户偏好学习的有效性,以及考虑用户响应偏好的广义需求侧资源响应优化的经济性。

关键词:

基金项目:

国家自然科学基金资助项目(51777126)。

通信作者:

作者简介:

孙伟卿(1985—),男,博士,副教授,博士生导师,主要研究方向:智能电网、电力系统储能、电力系统评估与优化。E-mail:sidswq@163.com
刘晓楠(1994—),女,通信作者,硕士研究生,主要研究方向:电力系统储能、电力市场。E-mail:lxn_1129@163.com
向威(1994—),男,硕士研究生,主要研究方向:电力系统储能、需求响应。E-mail:xw2013_usst@163.com


User Preference Learning and Response Optimization Based on Bayesian Inference
Author:
Affiliation:

1.School of Mechanical Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China;2.College of Electric Power Engineering, Shanghai University of Electric Power, Shanghai 200090, China

Abstract:

The flexible response aggregator (FRA) integrates user-side resources to participate in electricity market through demand response, which improves system reliability and economy and obtains revenue. However, consumer preference for electricity consumption modes and costs bring uncertainty to their response power quantity, which affects the accuracy of electricity declaration and market revenue of FRA. In this paper, the load resources involved in demand response are regarded as generalized demand side resources (GDSRs), and the user response preference model is constructed. Then, Bayesian inference is used to learn the user preference, and the probabilistic estimation of the response power quantity is obtained to generate the optimal response plan. Finally, the market revenue model of aggregator based on energy deviation penalty is established, and the user quotation and typical daily transaction data of PJM market in the United States are used for simulation. The simulation results verify the effectiveness of Bayesian inference for user preference learning and the economy of GDSR response optimization considering user response preference.

Keywords:

Foundation:
This work is supported by National Natural Science Foundation of China (No. 51777126).
引用本文
[1]孙伟卿,刘晓楠,向威,等.基于贝叶斯推断的用户偏好学习与响应优化[J].电力系统自动化,2020,44(19):92-100. DOI:10.7500/AEPS20200310002.
SUN Weiqing, LIU Xiaonan, XIANG Wei, et al. User Preference Learning and Response Optimization Based on Bayesian Inference[J]. Automation of Electric Power Systems, 2020, 44(19):92-100. DOI:10.7500/AEPS20200310002.
复制
分享
历史
  • 收稿日期:2020-03-10
  • 最后修改日期:2020-03-23
  • 录用日期:
  • 在线发布日期: 2020-10-15
  • 出版日期:
相关附件