The transition to electric vehicles (EVs) introduces new paradigms for emergency management, particularly in hurricane-prone regions. During extreme weather events, the vulnerability of the power grid becomes a critical bottleneck. Structural failures within the transmission and distribution networks can disable permanently installed EV charging stations precisely when they are most necessary for mass population evacuations.

The Research Problem

How can emergency managers systematically allocate limited mobile power resources to ensure EV operability when traditional grid-connected infrastructure fails?

Our recent study, “Robust Mobile Electric Vehicle Charging Solutions for Natural Disasters: A Multi-Criteria Resilience Analysis in Florida," formalizes a methodological framework to optimize the deployment of Mobile EV Charging Stations (MEVCS).

Data Collection & Preprocessing
Collect spatial & non-spatial data:
  • EV ownership by ZIP
  • Traffic intensity
  • Power grid fragility
  • Accessibility to shelters
  • Emergency route accessibility
Criteria Selection &
Normalization
Assign relative importance (based on stakeholder priorities) → Weight Assignment
Multi-Criteria Analysis
Normalize Evac Routes
Normalize Critical Facilities
Normalize Grid Failure
Normalize Traffic Volume
Normalize EV Density
Weighted Overlay Analysis
Suitability Surface
Optimal Location Selection
Extract High Suitability Areas
Generate Candidate Points
Select Distributed Locations
Final 14 Mobile Charging Station Locations

Figure 1: Methodological framework integrating spatial analytics and probabilistic power grid resilience modeling for MEVCS deployment.

Probabilistic Modeling of Grid Fragility

A novel contribution of this research is the explicit integration of power infrastructure vulnerability into the transportation planning model. Rather than assuming uniform charging operability, we employed a graph-based network model coupled with component-specific fragility curves.

The probability of transmission segment failure ($P_{failure}$) was calculated as a function of wind velocity ($V$) and the structural failure threshold ($d_{LS}$), followed by Monte Carlo simulations ($n=1000$) to evaluate system-wide cascading failures. Focusing on the Greater Tampa Bay area during Hurricane Ian (Category 5), our simulations under the most severe structural limit state ($d_{LS} = 0.245$) revealed that approximately 58.26% of existing charging stations would likely fail due to localized power generation and transmission line disruptions.

⚡
⚡
Network Status:
👆 Hover over the complex network above to simulate a focused hurricane strike. Watch the transmission lines in its path fail and observe the cascading power outages extending to nodes outside the path.

Multi-Criteria Decision Analysis (MCDA) via AHP

To determine optimal MEVCS placement, we utilized the Analytical Hierarchy Process (AHP) to derive criteria weights based on consensus from a panel of 12 domain experts across emergency management, power systems resilience, and urban planning. The normalized weights ($\omega$) were established as:

  • EV Ownership Density ($\omega_1 = 0.30$): Quantifying local charging demand using ZIP code registration counts.
  • Grid Failure Probability ($\omega_3 = 0.25$): Prioritizing areas with the highest simulated infrastructure fragility.
  • Evacuation Routes ($\omega_4 = 0.20$): Minimizing detour distances to designated evacuation corridors.
  • Critical Facilities ($\omega_5 = 0.15$): Ensuring service availability near hospitals and emergency shelters.
  • Traffic Volume ($\omega_2 = 0.10$): Using Annual Average Daily Traffic (AADT) to capture network usage intensity.

Validating Predictions with Satellite Imagery

To ensure the theoretical vulnerability model accurately reflected empirical conditions, we conducted an independent corroboration using satellite-derived nighttime light (NTL) imagery. By computing the normalized difference in nighttime radiance pre- and post-landfall, we observed distinct, localized decreases in brightness. These darkened zones aligned with high spatial correspondence to our Monte Carlo predicted outage clusters, validating that the framework successfully isolates the true geographic footprint of hurricane-induced disruptions.

🛰️
Satellite NTL View:
👆 Hover over the interactive map above to simulate the hurricane landfall and observe the power outages.

Strategic Deployment via Coverage-Aware Optimization

To translate the composite suitability surface into an actionable deployment strategy, we implemented a coverage-aware greedy selection algorithm. High-suitability areas were extracted and discretized into a 500-meter candidate grid.

To mitigate spatial clustering and maximize regional coverage, the algorithm iteratively selected the highest-ranked candidates while applying a 5-kilometer exclusion buffer. This process yielded 14 strategically distributed MEVCS locations that balance high empirical suitability with efficient spatial dispersion.

This methodology advances disaster operations research by explicitly coupling transportation network characteristics with power infrastructure resilience, providing decision-makers with a mathematically rigorous tool for emergency logistics planning.

Cite this paper

This research was published on Transport Geography.

@article{KHAYAMIM2026104739,
title = {Robust mobile electric vehicle charging solutions for natural disasters: A multi-criteria resilience analysis in Florida},
journal = {Journal of Transport Geography},
volume = {135},
pages = {104739},
year = {2026},
issn = {0966-6923},
doi = {https://doi.org/10.1016/j.jtrangeo.2026.104739},
url = {https://www.sciencedirect.com/science/article/pii/S0966692326001936},
author = {Razieh Khayamim and Mohammad Movahedi and Onur Alisan and Seçkin Özkul and Eren Erman Ozguven and Maxim A. Dulebenets},
}