面向多用户侧的火电机组负荷预测与储能调控算法  

Multi-user-side load forecasting and energy storage regulation algorithms for thermal power units

作　　者：岳昊 周毅 岳云力 武冰清 YUE Hao;ZHOU Yi;YUE Yunli;WU Bingqing(Economic and Technical Research Institute,State Grid Jibei Electric Power Co.,Ltd.,Beijing 100038,China)

机构地区：[1]国网冀北电力有限公司经济技术研究院,北京100038

出　　处：《电子设计工程》2025年第4期96-100,共5页Electronic Design Engineering

摘　　要：负荷预测对于电力生产、储能调控和输电网规划均具有重要意义。为保障电力生产过程中电能质量符合标准、储能调控策略及时有效,文中提出了一种面向多用户侧的火电机组负荷预测算法及储能调整框架。该方法采用K-means算法对用户负荷特性进行相似性聚类,刻画出具有不同用电习惯的用户种类,对不同用电习惯的用户采用Pyramid-CNN神经网络分类进行负荷预测并优化储能调控,根据电网中不同用户的实际占比情况对预测结果进行加权计算,从而得到最终的预测结果。以某电网用电数据为例进行的分析验证结果表明,所提负荷预测方法相比其他方法的预测准确度高、平均误差小,与实际情况的吻合度更高。Load forecasting is of great significance for power production,energy storage regulation,and transmission network planning.In order to ensure that the power quality meets the standards and the energy storage regulation strategy is timely and effective during the power production process,this paper proposes a multi user oriented load forecasting algorithm and energy storage adjustment framework for thermal power units.This method uses the K-means algorithm to perform similarity clustering on user load characteristics,characterizing different types of users with different electricity usage habits.Pyramid-CNN neural network classification is used for load prediction and optimization of energy storage regulation for users with different electricity usage habits.The prediction results are weighted based on the actual proportion of different users in the power grid to obtain the final prediction result.The analysis and validation results using electricity consumption data from a certain power grid show that the proposed load forecasting method has higher prediction accuracy,lower average error,and higher consistency with the actual situation compared to other methods.

关 键 词：聚类算法 负荷预测 神经网络 储能控制 电网规划 

分 类 号：TN713[电子电信—电路与系统] TM621[电气工程—电力系统及自动化]