考虑宏观政策的能源需求组合预测模型  

Combination Forecasting Model for Energy Demand Considering Micro Policy

作　　者：黄欣 吴杰康[2] 李红玲 李逸欣 郑敏嘉 李猛 吴伟杰 张伊宁 Huang Xin;Wu Jiekang;Li Hongling;Li Yixin;Zheng Minjia;Li Meng;Wu Weijie;Zhang Yining(Guangdong Power Grid Planning Research Center, Guangzhou 510000, Guangdong, China;School of Automation, Guangdong University of Technology, Guangzhou 510006, Guangdong, China)

机构地区：[1]广东电网规划研究中心,广东广州510000 [2]广东工业大学自动化学院,广东广州510006

出　　处：《四川电力技术》2021年第1期62-69,共8页Sichuan Electric Power Technology

基　　金：广东省科技计划项目(2020A050515003);广州市科技计划项目(202002030463);广东电网有限责任公司科技计划项目(037700KK52190004)。

摘　　要：针对目前能源需求预测考虑影响因素单一、忽视宏观政策影响因素、预测精度不足等问题,提出了一种融合灰色关联度分析和BP神经网络的能源需求组合预测方法。首先,根据宏观政策的针对性和方向性进行领域划分,分析各领域的指标;其次,利用灰色关联度分析法对分析后的各领域宏观政策初始指标进行关联度计算、排序、筛选;最后,将筛选出的初始指标作为BP神经网络的输入,利用神经网络预测能源需求,并进行实例仿真分析能源需求组合预测的预测结果。实例结果表明,所提出的能源需求组合预测方法着重考虑了宏观政策影响,且有效地提高了预测精度,具有实用性和可靠性。Aiming at the current problems of considering influence factors of energy demand forecasting in a single way,ignoring influence factors of macro policy and lacking in forecasting accuracy,a combined energy demand forecasting method together with grey correlation analysis and BP neural network is proposed.Firstly,the fields are divided according to the pertinence and direction of the macro policy,and then the indicators are analyzed in each field.Secondly,the grey correlation analysis method is used to calculate,sort and filter the initial indicators of the macro policy after the analysis.Finally,the selected initial indicators are used as the input of BP neural network,the neural network tools are used to achieve the purpose of energy demand forecasting,and a case simulation is performed to analyze the results of energy demand combination forecasting.The example results show that the proposed energy demand combination forecasting method focuses on the impact of macro policies,which effectively improves the forecasting accuracy and is practical and reliable.

关 键 词：能源需求预测 宏观政策 灰色关联度分析 BP神经网络 组合预测 

分 类 号：F206[经济管理—国民经济]