基于BP神经网络智能算法的园区综合能源系统负荷预测优化与工程实践  

Optimization and Practice of Load Forecasting for Integrated EnergySystem Based on BP Neural Network Intelligent Algorithm

作　　者：缪谦 MIAO Qian(Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China)

机构地区：[1]南京航空航天大学,南京210016

出　　处：《建筑节能(中英文)》2025年第4期83-88,139,共7页Building Energy Efficiency

摘　　要：在以往的传统能源系统投资、建设与营运管理中,不同类型的多能源系统之间缺乏耦合性的关系,各类能源系统往往缺少统一的、综合性的考量,一般均采取简单门类归集、总量调控,导致一个完整系统中各类能源冗余配置过度、全局构建合理性差、节能减排和绿色经济较差等后果。综合能源系统作为新兴概念,通过其有效的实施,可以整合水、电、气、热、冷、油等多种能源类别,实现多品质多类型能源的互补协调和优化,提升资源利用率和新能源消纳的比例,降低系统的整体投资、建设及运行成本。但是由于多类型能源的介入,导致了系统的不确定性与调控难度,如何提高系统在多能耦合下的供给和需求的协调统一,特别是负荷需求精准预测是亟需解决的问题。通过一种基于BP神经网络的方法开展园区综合能源系统负荷预测模型的研究,提出了一种遗传算法的智能优化方法,研究了园区综合能源各品质能源子系统间的耦合关系,并选取某示范案例进行了具体实践,验证了预测模型的优化有效性。为今后实施类似的综合能源工程提供了技术参考。In traditional energy system operation and management,there is a lack of coupling relationship between various energy systems,and there is no unified and comprehensive consideration for various types of energy.Often,simple category collection and total amount control are adopted,which will result in excessive redundant allocation of various energy sources in the system,poor layout rationality,and poor energy-saving and emission reduction effects.As an emerging concept,integrated energy systems can effectively integrate multiple energy categories such as water,electricity,gas,heat,cold,and oil through implementation,achieving complementary and coordinated optimization of multiple quality energy sources,effectively improving energy utilization efficiency,increasing the proportion of new energy consumption,and reducing system operating costs.However,the involvement of multiple types of energy in the system has increased the uncertainty and difficulty of regulation.How to improve the coordination and unity of supply and demand in the system under multi energy coupling,especially the accurate prediction of load demand,is an urgent problem that needs to be solved.The coupling relationship between various quality energy subsystems in a park’s comprehensive energy system is expounded,with a load forecasting model for the park’s comprehensive energy system based on BP neural network.A genetic algorithm intelligent optimization method is proposed,and a demonstration case of comprehensive energy in a park is selected for practice to verify the optimization effectiveness of the prediction model.Technical references are provided similar comprehensive energy projects in the future.

关 键 词：综合能源系统 负荷预测 多能协同 智能优化算法 

分 类 号：TU17[建筑科学—建筑理论] TK-9[动力工程及工程热物理]