摘要
随着可再生能源比例的增加和电网规模的不断扩大,电力系统的运行状态受不确定性因素的影响越来越大。以往研究中,以静态安全裕度为指标并考虑负荷延拓方向进行电力系统安全状态评估,未考虑在不同场景下或目标域数据集增大时的计算效率下降问题。因此,提出了一种基于迁移学习和卷积神经网络的电力系统安全状态评估方法。采用三层卷积层和三层池化层对所选取的负荷规律进行有效提取,并在不同场景下进行参数和权重迁移,实时快速掌握电力系统的运行状态。算例结果表明,该方法能在不同的场景保持较高的准确率,并且显著提高了计算效率。
With the increasing proportion of renewable energy and the continuous expansion of power grid scale,the operation state of power system is more and more affected by uncertain factors.In the previous studies,the static safety margin as an indicator and the load continuation direction were considered in the safety state assessment of power system,which did not involve the problem of computing efficiency decline in different scenarios or the increase of the target domain dataset.This paper proposed a safety state assessment method of power system based on transfer learning and convolution neural network.Three convolution layers and three pooling layers were used to effectively extract the selected load rules,and the operation state of the power system was grasped in real time.The simulation results show that the method can maintain a high accuracy in different scenarios and significantly improve the computing efficiency.
作者
吕琪
王长刚
王先伟
LV Qi;WANG Changgang;WANG Xianwei(Northeast Electric Power University,Jilin 132012,China;Key Laboratory of Modern Power System Simulation Control and New Green Power Technology of the Ministry of Education(Northeast Electric Power University),Jilin 132012,China)
出处
《电气应用》
2024年第9期30-36,共7页
Electrotechnical Application
基金
吉林省自然科学基金(YDZJ202101ZYTS149)。
关键词
电力系统
静态安全裕度
迁移学习
卷积神经网络
electric power system
static safety margin
transfer learning
convolutional neural network
