摘要
中国亟需在居民侧形成有效的市场化分时电价信号,以引导用户削峰填谷、促进新能源消纳,为保障电力系统安全稳定的经济运行提供支撑。基于此,利用居民用户智能电表电力消费数据并匹配大规模调查问卷数据,结合机器学习方法与经济学供需理论,构建异质性居民用户的电力需求曲线,提出一种考虑家庭个体异质性和用能行为时间异质性的量化方法,探究削峰、填谷等不同组合目标下的最优分时电价机制。仿真结果表明,居民需求侧响应有较大潜力,科学划分峰谷时段和合理确定峰谷电价价差可以有效缓解电力供需不匹配问题。夏季极端高温天气下,进一步加大尖峰时段和高峰时段间的电价差,可以有效缓解尖峰负荷压力。此外,建议在制定分时电价机制时,应关注居民家庭间和时段间的行为异质性以及自身电力供给结构。
China is in immediate need of implementing a robust market-driven time-of-use electricity pricing mechanism for residential consumers.This measure is crucial for guiding users in reducing electricity consumption during peak periods and increasing it during off-peak periods.Furthermore,it will facilitate the integration of renewable energy sources and contribute to the secure and stable economic functioning of the power system.This study employs smart meter electricity consumption data obtained from residential customers and integrates it with comprehensive survey questionnaire data on a broad scale.This approach integrates machine learning techniques with economic principles of supply and demand to develop power demand curves for diverse residential consumers.This study presents a quantitative approach that considers the diversity of individual households and the temporal variations in energy consumption patterns.It investigates the most effective time-of-use electricity pricing strategy for various objectives,such as reducing peak demand and filling off-peak periods.The findings from the simulation demonstrate a considerable opportunity for demand-side response among residential users.By employing a scientific approach to categorize peak and valley periods and establishing appropriate price differentials,it is possible to effectively address the imbalance between electricity supply and demand.In periods of exceptionally high temperatures throughout the summer season,implementing a greater price disparity between peak and off-peak hours can serve as an effective measure to mitigate the strain on the electrical grid during peak demand periods.Furthermore,the findings of this study indicate that when designing time-of-use pricing mechanisms,it is crucial to consider the diverse behavioral patterns exhibited by residential households and the variations across different time periods,as well as the unique characteristics of their energy supply systems.
作者
王博
张玉洁
刘向向
刘杰
李通
WANG Bo;ZHANG Yujie;LIU Xiangxiang;LIU Jie;LI Tong(School of Management and Economics,Beijing Institute of Technology,Beijing 100081,China;Power Supply Service Management Centre of State Grid Jiangxi Electric Power Co.,Ltd.,Nanchang Jiangxi 330000,China)
出处
《北京理工大学学报（社会科学版）》
北大核心
2023年第6期34-45,共12页
Journal of Beijing Institute of Technology：Social Sciences Edition
基金
国网江西省电力有限公司2021年科技项目(521852210017)。
关键词
需求侧管理
分时电价
异质性需求曲线
机器学习算法
低压台区
demand side management
time-of-use price
heterogeneous demand curve
machine learning algorithm
low-voltage distribution Area
