Scholarlink

<p>Most infrastructure flood-risk assessments overlook asset-level failure dynamics and the spatiotemporal cascades driven by interdependent networks. We develop a concise spatiotemporal framework that links flood hazards to multi-period failures by integrating remote-sensing-informed flood simulations across dry and wet seasons with a national charging-infrastructure network (43,162 assets across 12 UK regions). Disruption propagation is captured using a hybrid spatial–functional network that integrates node- and edge-level flood vulnerability features, coupled with a dynamic non-homogeneous Markov process to model multi-period failure transitions. We find: (1) station failure is primarily driven by network-level propagation, with wet-season co-amplification and dry-season edge dominance; edge attributes explain node failures more strongly than the reverse. (2) The risk network shows a dual-core small-world structure, where node failure correlates with degree, k-core, Katz, and strength (Spearman <i>r</i> ≈ 0.18–0.34), and edge failure correlates with endpoint-degree products (<i>r</i> ≈ 0.12–0.23). (3) Six propagation archetypes emerge, and future risk shifts from concentration in a few large communities toward dispersion across smaller clusters. The framework provides a clear, scalable tool for assessing systemic flood vulnerability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting multi period flood cascades and community failure in EV charging networks

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月7日

查看全文
  • Yunshan Wan,
  • Kaihan Zhang,
  • Rong Xia,
  • Zhixing Li,
  • Yecheng Zhang,
  • Paolo Vincenzo Genovese

摘要

Most infrastructure flood-risk assessments overlook asset-level failure dynamics and the spatiotemporal cascades driven by interdependent networks. We develop a concise spatiotemporal framework that links flood hazards to multi-period failures by integrating remote-sensing-informed flood simulations across dry and wet seasons with a national charging-infrastructure network (43,162 assets across 12 UK regions). Disruption propagation is captured using a hybrid spatial–functional network that integrates node- and edge-level flood vulnerability features, coupled with a dynamic non-homogeneous Markov process to model multi-period failure transitions. We find: (1) station failure is primarily driven by network-level propagation, with wet-season co-amplification and dry-season edge dominance; edge attributes explain node failures more strongly than the reverse. (2) The risk network shows a dual-core small-world structure, where node failure correlates with degree, k-core, Katz, and strength (Spearman r ≈ 0.18–0.34), and edge failure correlates with endpoint-degree products (r ≈ 0.12–0.23). (3) Six propagation archetypes emerge, and future risk shifts from concentration in a few large communities toward dispersion across smaller clusters. The framework provides a clear, scalable tool for assessing systemic flood vulnerability.