|
| 基于交通均衡理论与深度神经网络的电动汽车充电负荷预测 |
| Electric vehicle charging station load forecasting based on traffic equilibrium and deep neural networks |
| 投稿时间：2025-11-06 修订日期：2026-01-13 |
| DOI： |
| 中文关键词: 电动汽车 交通均衡 深度神经网络 充电负荷预测 特征选择 |
| 英文关键词: electric vehicle traffic equilibrium deep neural network charging load forecasting feature selection |
| 基金项目:国家电网有限公司科技项目 |
|
| 摘要点击次数: 0 |
| 全文下载次数: 0 |
| 中文摘要: |
| 电动汽车的大规模接入使配电网负荷呈现显著的时空波动特性，对电力系统的运行安全与充电设施规划带来挑战。传统时间序列和统计模型难以反映交通行为对充电负荷形成机制的影响。为此，提出一种基于交通均衡理论与深度神经网络相结合的电动汽车充电负荷预测方法。首先,建立考虑电量约束的交通均衡模型，模拟不同出行需求下的交通流量与充电行为，生成路网运行数据；随后，采用相关性分析与SHAP方法筛选关键特征，识别充电负荷的主要影响因素；最后，构建多层全连接神经网络，实现各充电站节点的充电负荷预测。结果显示，该模型具有良好的收敛性与稳定性，能够在精确预测充电负荷的同时，有效揭示交通行为与电力负荷间的耦合关系，为充电设施规划与配电系统优化提供技术支撑。 |
| 英文摘要: |
| The large-scale integration of electric vehicles (EVs) has caused significant spatial and temporal fluctuations in distribution net-work loads, posing challenges to power system operation secu-rity and charging infrastructure planning. Traditional time-series and statistical models fail to capture the influence of traffic be-havior on the formation of charging loads. To address this issue, a hybrid EV charging load forecasting method combining traffic equilibrium theory and deep neural networks (DNN) is pro-posed. First, a traffic equilibrium model considering energy con-straints is established to simulate traffic flow and charging be-havior under different travel demand scenarios, thereby generat-ing network operation data. Then, correlation analysis and SHAP methods are applied to identify key influencing features and determine the primary factors affecting charging loads. Finally, a multi-layer fully connected neural network is constructed to pre-dict charging loads at multiple station nodes. The results demon-strate that the proposed model exhibits good convergence and stability, and can accurately forecast charging loads while effec-tively revealing the coupling relationship between traffic behav-ior and power demand. This approach provides technical support for charging infrastructure planning and the optimization of modern distribution systems. |
|
View Fulltext
查看/发表评论 下载PDF阅读器 |
| 关闭 |