摘要
近年来,空调负荷的持续扩容赋予了需求侧更大的灵活性,但也给电网的安全稳定运行带来了一定的威胁。针对需求侧空调用户在“无感”情况下的调控及参与需求响应等问题,提出了一种基于深度强化学习技术的空调用户无感调控方法,并以广西南宁市的电力市场体系及气候特性数据为例,对该方法进行了验证计算。该强化学习方法以空调用户“无感”为约束,构建了用户基于不同目标参与调控的强化学习决策模型。将该空调调控方法应用于现实的数据案例中,其计算结果表明所提出的基于深度强化学习技术的空调用户无感调控方法可以代替空调用户自动进行空调设备的调度运行,从而在不影响用户舒适程度的前提下提升用户参与需求响应的意愿。
In recent years,the continuous expansion of air conditioning load has endowed the demand side with greater flexibility,but it has also posed a certain threat to the safe and stable operation of the power grid.To address the issues of demand-side air conditioning users'regulation and participation in demand response under"unperceived"conditions,a method based on deep reinforcement learning technology for unperceived regulation of air conditioning users is proposed.Taking the power market system and climate characteristic data of Nanning,Guangxi,as examples,this method has been verified through calculations.The reinforcement learning method,constrained by the air conditioning users'"unperceived"condition,constructs a reinforcement learning decision model for users to participate in regulation based on different objectives.Applying this air conditioning regulation method to real data cases,the calculation results show that the proposed method based on deep reinforcement learning technology can replace air conditioning users to automatically schedule the operation of air conditioning equipment,thereby enhancing the willingness of users to participate in demand response without affecting their comfort level.
作者
韩帅
卢健斌
吴宁
陈卫东
孙乐平
HAN Shuai;LU Jianbin;WU Ning;CHEN Weidong;SUN Leping(Electric Power Research Institute of Guangxi Power Grid Co.,Ltd.,Nanning 530023,China;Guangxi Key Laboratory of Intelligent Control and Maintenance of Power Equipment,Nanning 530023,China)
出处
《供用电》
北大核心
2024年第12期54-61,71,共9页
Distribution & Utilization
基金
中国南方电网有限责任公司科技项目(GXKJXM20220063)。
关键词
深度强化学习
无感调控
空调负荷
需求响应
响应意愿
deep reinforcement learning
senseless regulation
air conditioning load
demand response
willingness to response
