基于大规模冲击实验的居民电力需求响应行为机理、智能预测与管理策略研究

随着我国电气化水平不断提升以及城镇化进程推进，因家庭电力消费带来的能源消耗与排放问题日益严峻。居民电力需求响应作为“十四五”能源系统调度运行中的重要资源，成为深入贯彻落实国家能源安全战略的重要举措。本项目围绕居民需求响应行为，在充分考虑其影响因素基础上，针对典型场景系统设计开展大规模冲击实验，以此为基础系统集成多源数据构建居民需求响应数据库与标签库；识别影响居民需求响应行为关键影响因素、揭示作用机理；将因果分析框架与数据驱动的预测算法相结合，实现需求响应行为智能预测，为个性化需求响应管理策略制定提供科学依据。项目负责人与国家电力科学研究院、江西省电力有限公司建立了长期合作关系，2019年已经与合作单位开展7次夏季削峰场景下的大规模冲击实验，涵盖居民30余万户，采集家庭需求响应细粒度监测数据数据1.6亿条，为后续典型场景冲击试验开展提供借鉴，该项工作长期开展，确保数据获取与研究目标顺利完成。

With the promotion of national electrification process and the progress of urbanization, energy consumption and carbon emissions caused by household electricity consumption are getting increasingly serious. As an important resource in the in the scheduling and operation of the "14th five-year plan", residential power demand response has become an important measure to thoroughly implement the national energy security strategy. This project focuses on residents' demand response behavior and on the basis of fully considering the influencing factors of residents' demand response behavior, large scale controlled trials will be carried out with the systematic design of typical scenario, and high-frequency response behavior data will be collected. At the same time, multi-source data will be integrated to build the resident demand response database and tag library. On this basis, the key influencing factors affecting residents' demand response behavior will be identified and the mechanism of action will be revealed. The causal analysis framework is combined with data-driven prediction algorithm to realize intelligent prediction of demand response behavior, which provides scientific basis for formulating personalized demand response management strategy. A long-term partnership has been established among us and China Electric Power Research Institute and Jiangxi Electric Power Corporation, 7 large-scale controlled trials have been carried out under the summer peak reducing scenario in 2019 cover more than 300,000 households. 160 million fine-grained monitoring data of residential demand response have been collected, which provides reference for the following trials under other typical scenarios. This work have been included in Jiangxi Electric Power Corporation's long-term work plan, which ensure the data collection and successful completion of the research objectives.