| 含光储充的低压配电网净负荷区间预测方法 |
 点此下载全文 |
| 引用本文：崔威1,王俊龙1,王艺峰1,王 冰2,陈 磊2,董 斌3.含光储充的低压配电网净负荷区间预测方法[J].计算技术与自动化,2025,(1):118-124 |
| 摘要点击次数: 1027 |
| 全文下载次数: 216 |
|
|
| 中文摘要:准确的净负荷预测是有源低压配电网优化调度的依据。随着光伏渗透率的提高和储能、电动汽车充电站的普及，低压配电网净负荷的波动性和不确定性增大，常规确定性点负荷预测不能满足台区调度的需要。针对小数据集下净负荷不确定性预测问题,提出一种基于宽度学习的含光储充台区的净负荷区间预测方法。首先，通过变分模态分解算法将非平稳的净负荷数据分解为一系列相对平稳的模态分量，从而降低数据的复杂性；然后，这些模态分量通过适用于小规模数据集的宽度学习系统算法实现点预测；最后，在点预测的基础上建立一种区间优化模型，将点预测模型转换为区间预测模型。仿真实验表明，所提算法能够实现较精确的净负荷区间预测；与其他算法相比，在相同的置信区间要求下预测的区间更窄。 |
| 中文关键词:净负荷 区间预测 低压配电网 宽度学习系统 变分模态分解 |
| |
| Interval Prediction for Net Load in Low-Voltage Distribution Areas with Integrated Photovoltaic, Energy Storage, and Charging Station |
|
|
| Abstract:Accurate net load prediction serves as the foundation for optimal scheduling in active low-voltage distribution networks. With the increasing penetration of photovoltaic (PV) systems and the widespread adoption of energy storage and electric vehicle charging stations, the volatility and uncertainty of net load in low-voltage distribution networks have significantly increased. Traditional deterministic point load forecasting methods are insufficient to meet the scheduling needs of feeder-level management. To address the issue of net load uncertainty prediction in the context of small datasets, this study proposes an interval forecasting method for net load in feeder areas with integrated PV, storage, and charging systems, based on broad learning. First, the variational mode decomposition (VMD) algorithm is employed to decompose the non-stationary net load data into a series of relatively stationary modal components, thereby reducing data complexity. Subsequently, these modal components are subjected to point forecasting using the broad learning system (BLS) algorithm, which is well-suited for small-scale datasets. Finally, an interval optimization model is constructed based on the point forecasting results, transforming the point forecasting model into an interval forecasting model. Simulation experiments demonstrate that the proposed method achieves accurate interval forecasting of net load. Compared to other algorithms, the proposed method produces narrower prediction intervals under the same confidence level requirements. |
| keywords:net load interval prediction low-voltage distribution substation broad learning system variational mode decomposition |
| 查看全文 查看/发表评论 下载pdf阅读器 |
|
|
|
次