基于条件信息熵的电力系统行业用电数据因果性定量分析方法OA
Quantitative Causality Analysis of Industrial Electricity Consumption Data Based on Conditional Information Entropy
用电数据代表着电力系统用户负荷需求,其变化规律与人类现代社会生产生活存在着紧密联系.宏观、个体用电数据分别包含区域电力消费和个体用电行为等信息,对负荷预测、经济分析、电网运行具有重要意义;中观层面,行业用电数据所蕴藏的区域主导行业、产业链上下游结构等大量信息目前尚未得到充分挖掘.为此,该文提出一种基于条件信息熵的电力系统多元时序数据因果性定量分析方法,并将其用于行业用电量间关联关系分析中.基于某地2017年1月至2023年7月实际行业用电数据开展算例分析,对比所提方法与相关性方法在月度用电量预测中的应用效果,并基于数据成本-效用模型验证其有效性.该文所提方法能够量化分析行业用电数据中蕴含的有效信息价值,可为经济、社会发展提供一定参考.
Electricity consumption data not only represents the load demand for power users,but also shows a close connection between the production and residence in modern society.The macro and micro consumption data contain information about regional electricity demand and individual consumption behavior,which hold practical value in applications such as load forecast,power-economic analysis and system operation.However,information like dominant industry or industrial chain structure contained in sectoral electricity consumption data in intermediate perspective has not been fully revealed.To this end,a quantitative analysis method based on conditional information entropy for multi-time series data in power system is proposed and applied to analyze the causal relationship between industrial electricity consumption data.Case studies based on real data from Jan.2017 to Jul.2023 are conducted both qualitatively and quantitatively to validate the effectiveness of the proposed method.
张广伦;钟海旺
新型电力系统运行与控制全国重点实验室(清华大学电机工程与应用电子技术系),北京市 海淀区 100084新型电力系统运行与控制全国重点实验室(清华大学电机工程与应用电子技术系),北京市 海淀区 100084
信息技术与安全科学
信息熵因果分析行业用电量工业用电量
information entropycausal analysissectoral electricity consumptionindustrial power consumption
《中国电机工程学报》 2026 (10)
3953-3966,中插3,15
国家自然科学基金联合基金项目(U24B2077).Project Supported by the Joint Fund of National Natural Science Foundation of China(U24B2077).
评论