| 引用本文: | 陈逸飞,郑子萱,肖先勇,等.数据-物理混合驱动的配电网运行韧性评估方法与提升策略[J].电力系统保护与控制,2025,53(10):13-22.[点击复制] |
| CHEN Yifei,ZHENG Zixuan,XIAO Xianyong,et al.A data-physical hybrid-driven method for evaluating and enhancing the operational resilience of distribution networks[J].Power System Protection and Control,2025,53(10):13-22[点击复制] |
|
| 摘要: |
| 配电网多元敏感用户生产信息的不透明，使得电压暂降对敏感负荷的影响难以通过构建显性函数来准确量化，进一步限制了配电网运行韧性提升策略的有效性。为解决上述问题，提出了基于数据物理混合驱动的配电网运行韧性评估与提升方法。类比传统韧性指标及其定义构建了计及电压暂降对敏感用户影响的运行韧性指标。考虑到不同敏感用户对电压暂降的耐受特性不同，构建电压暂降轨迹特征体系以表征电压暂降下不同敏感负荷的响应特性，提出了数据驱动的配电网运行韧性评估模型。在此基础上，将数据驱动的韧性评估流程嵌入多目标储能优化配置的物理模型中。最后以IEEE33节点配电网为例进行算例分析。结果表明，所提数据物理混合驱动的储能优化配置模型能够解决电压暂降特征与运行韧性指标之间函数关系式难以显性表征的问题，能够在保障配电网运行经济性的同时改善运行韧性评估结果。 |
| 关键词: 配电网运行韧性 电压暂降 随机森林回归算法 储能优化配置 数据-物理混合模型 |
| DOI：10.19783/j.cnki.pspc.241043 |
| 投稿时间：2024-08-05修订日期：2024-11-08 |
| 基金项目:国家自然科学基金青年项目资助(52307128)；四川省科技计划资助(2023YFG0245) |
|
| A data-physical hybrid-driven method for evaluating and enhancing the operational resilience of distribution networks |
| CHEN Yifei1,ZHENG Zixuan1,XIAO Xianyong1,HU Wenxi1,CHEN Yunzhu1,WANG Yucai2 |
| (1. School of Electrical Engineering, Sichuan University, Chengdu 610065, China; 2. Ningdong Electric Power Supply
Company, State Grid Ningxia Electric Power Co., Ltd., Lingwu 750411, China) |
| Abstract: |
| The opacity of production information from various sensitive users in distribution networks makes it difficult to construct explicit functions to accurately quantify the impact of voltage sags on sensitive loads, thereby limiting the effectiveness of resilience enhancement strategies for distribution network operations. To address this issue, a data-physical hybrid-driven method is proposed for evaluating and enhancing distribution network operational resilience. By drawing analogies with traditional resilience indices and their definitions, a new operational resilience index that accounts for the impact of voltage sags on sensitive users is constructed. Considering the different tolerance characteristics of various sensitive users to voltage sags, a voltage sag trajectory characteristic system is established to represent the response characteristics of different sensitive loads. A data-driven model is then proposed to evaluate distribution network operational resilience. On this basis, the data-driven resilience evaluation process is embedded into a physical model for multi-objective energy storage optimization. Finally, a case study based on the IEEE33-bus distribution network is conducted. The results demonstrate that the proposed data-physical hybrid-driven model for energy storage optimization can address the challenge of explicitly modelling the functional relationship between voltage sag characteristics and resilience indices, and can enhance resilience evaluation results while ensuring economic operation of the distribution network. |
| Key words: distribution network operational resilience voltage sag random forest regression (RFR) algorithm energy storage optimization configuration data-physical hybrid model |
