| 引用本文: | 杨桂兴,王维庆,姚红雨,郭小龙,袁铁江.面向广域分布的家庭蓄热式电采暖集群控制方法[J].电力自动化设备,2023,43(8):56-62 |
| YANG Guixing,WANG Weiqing,YAO Hongyu,GUO Xiaolong,YUAN Tiejiang.Control method of wide-area distributed household regenerative electric heating clusters[J].Electric Power Automation Equipment,2023,43(8):56-62 |
|
| 摘要: |
| 针对面向新能源消纳调制广域分布的家庭蓄热式电采暖(REH)负荷的时空分布特征存在维度大、调控难的问题，提出了一种基于历史数据的REH控制方法。根据用户房屋参数、REH类型等信息对REH用户进行聚类以降低策略的求解维度；考虑用户舒适度、REH运行约束条件，并根据不同的控制目标建立对应的目标函数，采用粒子群优化算法优化历史数据中各REH集群的用电策略；基于卷积神经网络学习状态特征与REH用电策略之间的关系，利用该卷积神经网络生成REH的实时运行策略。基于某地区冬季电网的发电数据，采用蒙特卡罗方法模拟10000台REH的用电需求并进行仿真分析，以验证所提方法的有效性。结果表明所提方法能在满足策略生成时效性以及用户舒适度的前提下，有效促进新能源消纳，并减少用户的取暖费用，平抑负荷波动。 |
| 关键词: 柔性负荷 蓄热式电采暖 卷积神经网络 粒子群优化算法 |
| DOI：10.16081/j.epae.202303009 |
| 分类号:TM73 |
| 基金项目:国家自然科学基金资助项目(52267005) |
|
| Control method of wide-area distributed household regenerative electric heating clusters |
|
YANG Guixing1,2, WANG Weiqing1, YAO Hongyu3, GUO Xiaolong2, YUAN Tiejiang3
|
|
1.School of Electrical Engineering, Xinjiang University, Urumqi 830049, China;2.State Grid Xinjiang Electric Power Co.,Ltd.,Urumqi 830049, China;3.School of Electrical Engineering, Dalian University of Technology, Dalian 116081, China
|
| Abstract: |
| Aiming at the problems of large dimension and difficult regulation of the spatial and temporal distribution characteristics of wide-area distributed household regenerative electric heating(REH) load for new energy consumption modulation, a REH control method based on historical data is proposed. According to the information such as user house parameters, REH type, and so on, the REH users are clustered to reduce the solving dimension of the strategy. Considering the user comfort and the REH operation constraints, the corresponding objective functions are established according to different control objectives. The particle swarm optimization algorithm is used to optimize the historical power consumption strategy of each REH cluster. The relationship between the learning state characteristics and the power consumption strategy of REH is studied by using the convolutional neural network, and the real-time operation strategy of REH is generated based on this convolutional neural network. Based on the power generation data of a region in winter, Monte Carlo method is used to simulate the power demand of 10000 REH units and the simulation analysis is carried out to verify the effectiveness of the proposed method. The results show that the proposed method can effectively promote the consumption of new energy, reduce the heating cost of users and smooth load fluctuation under the premise of meeting the timeliness of strategy generation and user comfort. |
| Key words: flexible load regenerative electric heating convolutional neural network particle swarm optimization algorithm |
