基于数据驱动的家庭能量实时经济调控方法  

Data-driven Real-time Economic Regulation Method for Household Energy

作　　者：王玉彬 董伟 陈源奕 杨强[1] WANG Yubin;DONG Wei;CHEN Yuanyi;YANG Qiang(College of Elctrical Engineering,Zhejiang University,Hangzhou 310027,China;Polyechnic Institute,Zhejang University,Hangzhou 310058,China)

机构地区：[1]浙江大学电气工程学院,浙江省杭州市310027 [2]浙江大学工程师学院,浙江省杭州市310058

出　　处：《电力系统自动化》2022年第13期21-29,共9页Automation of Electric Power Systems

基　　金：国家自然科学基金资助项目(52177119)。

摘　　要：分布式电源和大量用户侧可调资源的接入使得通过需求侧响应提高系统的用能效率、节约用电成本以及实现清洁能源转型成为现实。然而,分布式发电的不确定性和间歇性以及用户负荷的随机波动使得依赖于预测数据的模型驱动调度方法存在预测误差导致调度效果不佳的问题。针对以上问题,提出了一种基于数据驱动的家庭能量实时经济调控方法。该方法首先建立家庭能量系统的数学模型,然后利用历史数据基于模型驱动方法构建训练数据集;所构建的训练数据集将用于人工神经网络的监督学习,从而建立基于数据驱动的调度决策模型;之后,在新的场景到来时由该模型输出调度结果并施加设备约束得到调度决策。仿真表明,所述方法可以基于电价变化协调用电器和储能系统的运行,在保证用电器和储能系统安全运行的前提下实现经济运行。The integration of distributed generators and a massive amount of customer-side adjustable resources makes it possible to improve system energy efficiency, save electricity costs and achieve clean energy transition through demand-side response.However, the uncertainty and intermittency of distributed generation as well as the random fluctuation of customer load make the model-driven regulation methods that rely on forecasted data have the problems of poor regulation results due to forecast errors. To address the above problems, a data-driven real-time economic regulation method for the household energy is proposed. The method first establishes a mathematical model of the household energy system, and then uses historical data to construct a training sample dataset based on the model-driven approach. The constructed training dataset will be used for the supervised learning of the artificial neural network to build a data-driven regulation decision model. Afterwards, when a new scenario arrives, the model outputs the regulation results and imposes equipment constraints to obtain regulation decisions. Simulations demonstrate that the proposed method can coordinate the operation of the electrical appliances and the energy storage system based on the change of electricity price, which can achieve the economic operation while ensuring the safe operation of the electrical appliances and the energy storage system.

关 键 词：家庭能量系统 经济调控 人工神经网络 数据驱动 设备约束 

分 类 号：TM73[电气工程—电力系统及自动化]