| 引用本文: | 王维皓,马瑞,石庆鑫,刘文霞,杨育超,张晓晨.基于元学习的城市电-水耦合网络脆弱性评估方法[J].电力自动化设备,2026,46(1):23-30，70 |
| WANG Weihao,MA Rui,SHI Qingxin,LIU Wenxia,YANG Yuchao,ZHANG Xiaochen.Vulnerability assessment method of urban electricity-water coupling network based on meta-learning[J].Electric Power Automation Equipment,2026,46(1):23-30，70 |
|
| 摘要: |
| 为评估电网失负荷导致的供水服务中断对城市关键基础设施系统脆弱性的影响，并解决故障传导建模中的水力计算难题，提出一种基于元学习框架的城市电-水耦合网络脆弱性评估方法。基于系统动力学理论与网络拓扑特征，构建电力-供水物理耦合网络的3层故障传播模型，并建立电网、水网潮流计算模型。针对传统水力计算模型存在的参数不精确以及数据难以获取的问题，引入元学习方法进行水力计算，实现电网中断对供水网络影响的评估。以某城市电网及供水网为研究对象，验证了所提方法的水力计算精度，所提方法可降低样本需求，显著提升模型计算效率。通过构建包含供水服务损失度、地块功能重要度的多维脆弱性指标，在N-k极端攻击场景下识别出电网关键脆弱节点，这些节点失效将导致城市供水服务能力下降34 %以上。 |
| 关键词: 城市电网 城市供水网 电-水耦合 故障传导 元学习 脆弱性 韧性 |
| DOI：10.16081/j.epae.202511011 |
| 分类号:TM73 |
| 基金项目:国家自然科学基金资助项目(52307094) |
|
| Vulnerability assessment method of urban electricity-water coupling network based on meta-learning |
|
WANG Weihao1, MA Rui1, SHI Qingxin1, LIU Wenxia1, YANG Yuchao1, ZHANG Xiaochen2
|
|
1.School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China;2.Chaoyang Power Supply Company of State Grid Beijing Electric Power Company, Beijing 100020, China
|
| Abstract: |
| In order to assess the impact of water service outage caused by power grid lost load on the vulnerability of urban critical infrastructure system and solve the hydraulic calculation problem in fault propagation modeling, a vulnerability assessment method for urban electricity-water coupling network is proposed based on a meta-learning framework. Based on system dynamic theory and network topology characteristic, a three-layer fault propagation model for electricity-water supply physically coupling network is constructed, and the power flow and water flow calculation models for power grid and water network are built. Aiming at the problems that the parameters are imprecise and the data is hard to acquire in the traditional hydraulic calculation models, a meta-learning method is introduced for hydraulic calculation, and the impact of power grid outage on water supply network is realized. The power grid and water supply network in a city are taken as research objects, the hydraulic calculation accuracy of the proposed method is verified, the proposed method can reduce the sample requirement and significantly improve the model calculation efficiency. The critical vulnerable nodes in power grid are identified under N-k extreme attack scenarios by constructing a multi-dimensional vulnerability index incorporating the loss degree of water supply service and the importance of land function, the failure of these nodes will cause water supply service capacity of the city to decline by over 34%. |
| Key words: urban power grid urban water supply network electricity-water coupling fault propagation meta-learning vulnerability resilience |
