“双碳”背景下结合系统动力学和弹性网络回归的区域电力需求中长期预测方法  

A mid-long regional electricity demand forecasting method based on system dynamics and elastic network regression under dual carbon targets

作　　者：游铭豪 顾洁[1] 储琳琳 宗明 YOU Minghao;GU Jie;CHU Linlin;ZONG Ming(Department of Electrical Engineering and Automation,School of Electronic Information and Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;State Grid Shanghai Electric Power Company,Shinan Electric Power Company,Shanghai 200030,China)

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240 [2]国网上海市电力公司市南供电公司,上海200030

出　　处：《供用电》2024年第7期72-83,共12页Distribution & Utilization

基　　金：国家自然科学基金项目(52307119);国网上海市电力公司科技项目(SGSHSN00ZSJS2202767)。

摘　　要：在“双碳”目标的驱动下,地区社会经济发展、能源生产与消费结构将迎来深刻的变革,从而引发区域电力需求预测影响因素和变化趋势的新特征。作为实现“双碳”目标的关键领域,电力行业亟须深入探索新形势下的区域电力需求预测方法。针对这一挑战,提出了一种创新性的区域电力需求预测方法,该方法融合了系统动力学(system dynamics,SD)与弹性网络回归分析的优势。综合考虑了经济、政策、能源等多重影响因素,并基于SD模型建立了一个包含社会经济、清洁能源发电占比、电气化率、能源强度四大子系统的电力需求场景生成模型。为了进一步提升模型的预测精度,引入了弹性网络回归方法,提升了模型的拟合效果。以我国某地区为例,基于SD模型仿真结果对该地区“双碳”背景下的不同情景电力需求进行了预测,并与传统的电力需求预测方法进行了对比。结果表明所提出的预测方法能够更准确地反映区域电力需求的动态特征,为实现“双碳”目标提供了科学的决策支持。Driven by the dual carbon goals,significant changes will occur in socio-economic development,energy production,and consumption structure,leading to new characteristics in the influencing factors and trends of regional electricity demand forecasting.The power industry is a key area to ensure the practical realization of the dual carbon targets,and it is necessary to conduct research on regional electricity demand forecasting methods under the new situation.A regional electricity demand prediction method based on System Dynamics(SD)and elastic network regression is proposed to address the issue of electricity demand prediction under dual carbon targets.Firstly,a power demand scenario generation model was established by considering various influencing factors such as economy,policies,and energy,including four subsystems:socio-economic,electrification rate,energy intensity,and the proportion of clean energy in power generation.Then,by introducing the elastic network regression method,the fitting effect of the model is improved,and a more accurate regional electricity demand forecasting model is constructed.Taking a certain region in China as an example,the SD model simulation results proposed in this paper are used to predict the electricity demand in different scenarios under the background of dual carbon in the region,and compared with traditional electricity demand forecasting methods.The results showed that the prediction method proposed in this article can more accurately forecast electricity demand in the region,providing decision-making support for the dual carbon targets.

关 键 词：电力需求预测 “双碳”目标 弹性网络回归 系统动力学 场景生成 

分 类 号：TM715[电气工程—电力系统及自动化]