pith. sign in

arxiv: 2607.00585 · v1 · pith:TYRSWW5Nnew · submitted 2026-07-01 · 📡 eess.SY · cs.SY

Mobility Safe Adaptive Reserve Certification for Electric Vehicle Hydrogen Bus and Building Resilience Hubs

Yifan Wang This is my paper

Pith reviewed 2026-07-02 07:45 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords electric vehicleshydrogen busesresilience hubsconformal inferencereserve certificationmobility safetybuilding load modelingoutage scenarios
0
0 comments X

The pith

A mobility-first certification policy achieves 100% bus protection and commitment delivery while serving 20.5% of critical-building demand in zero-emission hubs.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that shared hubs for EV charging, hydrogen buses, and building backup cannot rely on raw export capacity because hydrogen exports can strand buses and demand varies seasonally. It develops a framework that pairs a physics-hybrid building-load model with conformal calibration methods and a scheduling rule that protects bus service first. On 66,816 held-out outage scenarios drawn from real EV sessions, bus schedules, and weather data, only this mobility-first policy reached full commitment delivery, full bus protection, and zero mean hydrogen shortfall. Other policies either failed to protect buses or delivered far fewer commitments. The result indicates that resilience in these hubs hinges on service-aware reserve certification rather than maximum export volume.

Core claim

The certified mobility-first policy was the only tested policy to achieve 100% commitment delivery, 100% bus protection, and zero mean bus-hydrogen shortfall, while serving 20.5% of critical-building demand. Under seasonal load shifts, adaptive conformal inference raised late-period coverage from 0.687 to 0.831 and reached 0.891 overall against a 0.90 target with lower mean reserve than static methods.

What carries the argument

The mobility-safe reserve certification framework, which integrates a physics-hybrid universal differential equation building-load twin, one-sided split conformal reserve calibration, adaptive conformal inference for seasonal drift, and a mobility-first scheduling rule that protects post-event bus service before assigning hydrogen to buildings.

If this is right

  • A mobility-blind hydrogen-export policy served 39.2% of building demand but protected buses in 0% of cases and caused a 426.7 kg mean bus-hydrogen shortfall.
  • A nominal mean-resource promise delivered only 45.4% of commitments.
  • Adaptive conformal inference kept low late-coverage variability and the lowest mean reserve across 12 building and seed drift runs.
  • The framework was evaluated on 495,221 real EV sessions, GTFS-derived bus service days, and EnergyPlus simulations under TMY3 weather.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same priority-rule approach could be tested on microgrids that serve both transportation fleets and stationary loads.
  • Live integration of bus schedule data might further tighten the reserve calibration in operational settings.
  • Extending the building-load twin to additional climate zones would test whether the coverage guarantees hold under wider weather variability.

Load-bearing premise

The physics-hybrid universal differential equation building-load twin combined with one-sided split conformal calibration and adaptive conformal inference produces valid coverage and scheduling outcomes on the held-out outage scenarios.

What would settle it

Apply the same four policies to a fresh collection of outage scenarios generated from different weather years or regions and check whether the mobility-first policy still records 100% bus protection and zero mean shortfall.

Figures

Figures reproduced from arXiv: 2607.00585 by Yifan Wang.

Figure 1
Figure 1. Figure 1: Planned framework schematic for the EV–hydrogen-bus–building resilience hub. [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Data and system evidence for the coupled hub. The figure summarizes EV [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Operational tradeoff between apparent building service and mobility-safe deliv [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Forecasting and reserve safety. Panel (a) compares the physics-hybrid UDE, [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
read the original abstract

Zero-emission mobility depots are becoming resilience assets because one site can host EV charging, hydrogen-bus operation, stationary conversion equipment, and nearby critical-building backup. The key question is not raw outage export capacity: hydrogen exported to buildings can strand buses, EV availability is stochastic, and building demand shifts seasonally. We introduce a mobility-safe reserve certification framework for a coupled EV, hydrogen-bus, and critical-building hub. It combines a physics-hybrid universal differential equation building-load twin, one-sided split conformal reserve calibration, adaptive conformal inference for seasonal drift, and a mobility-first scheduling rule that protects post-event bus service before assigning hydrogen to buildings. Evaluation uses 495,221 real EV charging sessions across eight regions, AC Transit GTFS-derived hydrogen-bus service days, and EnergyPlus 25.2 simulations under real TMY3 weather. Across 66,816 held-out outage scenarios, a mobility-blind hydrogen-export policy served 39.2\% of building demand but protected buses in 0\% of cases and caused a 426.7 kg mean bus-hydrogen shortfall. A nominal mean-resource promise delivered only 45.4\% of commitments. The certified mobility-first policy was the only tested policy to achieve 100\% commitment delivery, 100\% bus protection, and zero mean bus-hydrogen shortfall, while serving 20.5\% of critical-building demand. Under a summer-to-winter load shift, adaptive conformal inference raised late-period empirical coverage from 0.687 to 0.831 and reached 0.891 overall coverage against a 0.90 target with lower mean reserve than static split conformal. Across 12 building/seed drift runs, it kept low late-coverage variability and the lowest mean reserve. These results show that resilience value in shared zero-emission hubs depends on service-aware certification, not raw export capacity alone.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 3 minor

Summary. The manuscript introduces a mobility-safe reserve certification framework for coupled EV charging, hydrogen-bus, and critical-building resilience hubs. It combines a physics-hybrid universal differential equation (UDE) building-load twin, one-sided split conformal reserve calibration, adaptive conformal inference to handle seasonal drift, and a mobility-first scheduling rule that prioritizes post-event bus service. On 495,221 real EV sessions, GTFS-derived bus schedules, EnergyPlus simulations, and 66,816 held-out outage scenarios, the certified mobility-first policy is reported as the only tested approach to achieve 100% commitment delivery, 100% bus protection, and zero mean bus-hydrogen shortfall while serving 20.5% of building demand; adaptive conformal inference is shown to raise late-period empirical coverage from 0.687 to 0.831 (overall 0.891 vs. 0.90 target) with lower mean reserve than static calibration.

Significance. If the coverage properties hold, the work provides concrete evidence that service-aware certification, rather than raw export capacity, determines resilience value in shared zero-emission hubs. Strengths include the scale of the held-out evaluation (66,816 scenarios), use of real multi-region EV data and TMY3 weather, and direct policy comparisons that quantify trade-offs between building service and bus protection. The empirical demonstration of adaptive conformal inference under documented summer-to-winter load shift is a useful data point for non-stationary conformal methods in energy systems.

major comments (2)
  1. [§4.2] §4.2 (Adaptive Conformal Inference for seasonal drift): the manuscript applies adaptive conformal inference to correct for summer-to-winter building-load drift induced by the UDE twin but provides neither a derivation nor a cited reference establishing that the procedure retains finite-sample marginal coverage under the resulting non-exchangeable process. The reported empirical coverage (late-period 0.831, overall 0.891) is therefore insufficient to underwrite the 100% bus-protection and zero-shortfall claims, which rest on the validity of the certified reserves.
  2. [§3.1 and §5.1] §3.1 and §5.1 (one-sided split conformal calibration + mobility-first rule): the definition of the certified reserve and the mobility-first scheduling rule are presented without an explicit statement of the exchangeability assumption required for the one-sided conformal guarantee, nor a sensitivity analysis showing how violation of that assumption (via UDE-induced non-stationarity) would affect the 100% commitment-delivery metric on the held-out set.
minor comments (3)
  1. [Table 2] Table 2: the column headers for the four policies are not fully aligned with the policy descriptions in §4.3, making it difficult to map the 39.2% building-service figure to the mobility-blind policy.
  2. [§2.3] §2.3: the UDE building-load twin is introduced with a hybrid physics-data loss but the precise weighting between the physics residual and data-fitting terms is not stated, complicating reproduction of the twin used for the 66,816 scenarios.
  3. [Figure 4] Figure 4 caption: the legend for the adaptive vs. static conformal curves does not indicate whether the shaded bands represent standard deviation across the 12 building/seed runs or a different quantity.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for these precise comments on the theoretical grounding of our conformal methods. We address each point below with proposed revisions.

read point-by-point responses

  1. Referee: [§4.2] §4.2 (Adaptive Conformal Inference for seasonal drift): the manuscript applies adaptive conformal inference to correct for summer-to-winter building-load drift induced by the UDE twin but provides neither a derivation nor a cited reference establishing that the procedure retains finite-sample marginal coverage under the resulting non-exchangeable process. The reported empirical coverage (late-period 0.831, overall 0.891) is therefore insufficient to underwrite the 100% bus-protection and zero-shortfall claims, which rest on the validity of the certified reserves.

    Authors: We agree that no derivation or reference for finite-sample marginal coverage under the induced non-exchangeability is provided. The 100% bus-protection, commitment-delivery, and zero-shortfall figures are strictly empirical results measured on the 66,816 held-out outage scenarios after applying the adaptively calibrated reserves inside the mobility-first scheduler. We will revise §4.2 to cite representative adaptive conformal literature for non-stationary settings and to state explicitly that the coverage claims rest on observed empirical performance rather than a proven finite-sample guarantee. revision: partial

  2. Referee: [§3.1 and §5.1] §3.1 and §5.1 (one-sided split conformal calibration + mobility-first rule): the definition of the certified reserve and the mobility-first scheduling rule are presented without an explicit statement of the exchangeability assumption required for the one-sided conformal guarantee, nor a sensitivity analysis showing how violation of that assumption (via UDE-induced non-stationarity) would affect the 100% commitment-delivery metric on the held-out set.

    Authors: We will insert an explicit statement of the exchangeability assumption in §3.1. In §5.1 we will add a sensitivity subsection that re-evaluates commitment-delivery rates on the held-out set under controlled increases in non-stationarity (by scaling the UDE drift parameters), thereby quantifying robustness of the 100% metric. revision: yes

standing simulated objections not resolved
  • Lack of a derivation establishing finite-sample marginal coverage for adaptive conformal inference under the non-exchangeable process induced by the UDE twin.

Circularity Check

0 steps flagged

No circularity: empirical evaluation on held-out scenarios is independent of calibration inputs

full rationale

The paper describes a composite framework (physics-hybrid UDE twin + one-sided split conformal calibration + adaptive conformal inference + mobility-first rule) and reports direct performance metrics (100% commitment delivery, 100% bus protection, zero mean shortfall) on 66,816 explicitly held-out outage scenarios drawn from real EV sessions, GTFS data, and EnergyPlus simulations. No equation or procedure is shown reducing the reported outcomes to quantities fitted on the evaluation distribution itself; the adaptive coverage numbers (0.831 late-period, 0.891 overall) are presented as empirical observations against a 0.90 target rather than as theoretically guaranteed or self-referential quantities. The derivation chain therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no explicit free parameters, axioms, or invented entities; no details on model fitting, assumptions in conformal calibration, or new postulated quantities are available.

pith-pipeline@v0.9.1-grok · 5868 in / 1045 out tokens · 32672 ms · 2026-07-02T07:45:31.989296+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages · 1 internal anchor

  1. [1]

    W. Tian, P. Talebizadehsardari, Energy cost and efficiency analysis of building resilience against power outage by shared parking station for electric vehicles and demand response program, Energy (2021).doi: 10.1016/j.energy.2020.119058

    work page doi:10.1016/j.energy.2020.119058 2021

  2. [2]

    H. Liu, A. Abdin, J. Puchinger, Improving critical buildings energy resilience via shared autonomous electric vehicles: A sequential opti- mization framework, Computers & Operations Research (2024).doi: 10.1016/j.cor.2023.106513

    work page doi:10.1016/j.cor.2023.106513 2024

  3. [3]

    J. A. Manzolli, S. Yu, D. D’Apice, L. F. Miranda-Moreno, Bal- ancing energy resilience and mobility: a multi-objective strategy for deploying shared autonomous electric vehicles during power out- ages, npj Sustainable Mobility and Transport (2026).doi:10.1038/ s44333-026-00081-9

    work page 2026

  4. [4]

    H. Liu, A. Vlachokostas, E. Kontou, Resilience and environmental ben- efits of electric school buses as backup power for educational func- tions continuation during outages, Building and Environment (2025). doi:10.1016/j.buildenv.2024.112329

    work page doi:10.1016/j.buildenv.2024.112329 2025

  5. [5]

    Y. Jia, K. An, Multi-depot battery electric bus scheduling and charging coordination under resource limitations, Applied Energy (2025).doi: 10.1016/j.apenergy.2025.126444. 15

    work page doi:10.1016/j.apenergy.2025.126444 2025

  6. [6]

    Zhang, Z

    X. Zhang, Z. Hu, Y. Cao, A. M. Y. M. Ghias, Novel data-driven de- centralized coordination model for electric vehicle aggregator and en- ergy hub entities in multi-energy system using an improved multi-agent drl approach, Applied Energy (2023).doi:10.1016/j.apenergy.2023. 120902

    work page doi:10.1016/j.apenergy.2023 2023

  7. [7]

    M. F. Tahir, J. Hu, A. Khan, J. Zhu, Sustainable hybrid station de- sign framework for electric vehicle charging and hydrogen vehicle refu- eling based on multiple attributes, Energy Conversion and Management (2024).doi:10.1016/j.enconman.2023.117922

    work page doi:10.1016/j.enconman.2023.117922 2024

  8. [8]

    Kazemi, T

    S. Kazemi, T. Khalili, S. Beydaghi, M. Moeini-Aghtaie, Hydrogen vehi- cle operational management in the context of green-hydrogen energy hub applications, International Journal of Hydrogen Energy (2025). doi:10.1016/j.ijhydene.2025.152359

    work page doi:10.1016/j.ijhydene.2025.152359 2025

  9. [9]

    Haddad, S

    R. Haddad, S. Asadi, N. Alemazkoor, Integrated sizing and management of residential energy systems for electric and hydrogen vehicle charging, Journal of Building Engineering (2025).doi:10.1016/j.jobe.2025. 114675

    work page doi:10.1016/j.jobe.2025 2025

  10. [10]

    Universal Differential Equations for Scientific Machine Learning

    C. Rackauckas, Y. Ma, J. Martensen, C. Warner, K. Zubov, R. Supekar, D. Skinner, A. Ramadhan, A. Edelman, Universal differential equations for scientific machine learning (2020).arXiv:2001.04385. URLhttps://arxiv.org/abs/2001.04385

    work page internal anchor Pith review Pith/arXiv arXiv 2020

  11. [11]

    Gokhale, B

    G. Gokhale, B. Claessens, C. Develder, Physics informed neural net- works for control oriented thermal modeling of buildings, Applied En- ergy (2022).doi:10.1016/j.apenergy.2022.118852

    work page doi:10.1016/j.apenergy.2022.118852 2022

  12. [12]

    Raissi, P

    M. Raissi, P. Perdikaris, G. E. Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of ComputationalPhysics378(2019)686–707.doi:10.1016/j.jcp.2018. 10.045

    work page doi:10.1016/j.jcp.2018 2019

  13. [13]

    Shafer, V

    G. Shafer, V. Vovk, A tutorial on conformal prediction, Journal of Ma- chine Learning Research 9 (2008) 371–421. URLhttps://jmlr.org/papers/v9/shafer08a.html 16

    work page 2008

  14. [14]

    J. Lei, M. G’Sell, A. Rinaldo, R. J. Tibshirani, L. Wasserman, Distribution-free predictive inference for regression, Journal of the American Statistical Association 113 (523) (2018) 1094–1111.doi: 10.1080/01621459.2017.1307116

    work page doi:10.1080/01621459.2017.1307116 2018

  15. [15]

    M.Borrotti, Quantifyinguncertaintywithconformalpredictionforheat- ing and cooling load forecasting in building performance simulation, En- ergies 17 (17) (2024) 4348.doi:10.3390/en17174348

    work page doi:10.3390/en17174348 2024

  16. [16]

    Gibbs, E

    I. Gibbs, E. Candès, Adaptive conformal inference under distribution shift, in: Advances in Neural Information Processing Systems, Vol. 34, 2021, pp. 1660–1672. URLhttps://proceedings.neurips.cc/paper/2021/hash/ 0d441de75945e5acbc865406fc9a2559-Abstract.html

    work page 2021

  17. [17]

    Suresh, A

    V. Suresh, A. K. Swain, S. Revathi, J. M. Guerrero, Mamba based adaptive conformal inference for probabilistic short-term load forecast- ing, Knowledge-Based Systems (2025).doi:10.1016/j.knosys.2025. 114222

    work page doi:10.1016/j.knosys.2025 2025

  18. [18]

    URLhttps://data.townofcary.org/

    Town of Cary, Town of cary open data (2026). URLhttps://data.townofcary.org/

    work page 2026

  19. [19]

    Amara, Ev load open data (2026)

    Y. Amara, Ev load open data (2026). URLhttps://github.com/yvenn-amara/ev-load-open-data

    work page 2026

  20. [20]

    URLhttps://www.actransit.org/data-api-resource-center

    Alameda-Contra Costa Transit District, Data api resource center (2026). URLhttps://www.actransit.org/data-api-resource-center

    work page 2026

  21. [21]

    D. B. Crawley, L. K. Lawrie, F. C. Winkelmann, W. F. Buhl, Y. J. Huang, C. O. Pedersen, R. K. Strand, R. J. Liesen, D. E. Fisher, M. J. Witte, J. Glazer, Energyplus: creating a new-generation building energy simulation program, Energy and Buildings 33 (4) (2001) 319–331.doi: 10.1016/S0378-7788(00)00114-6

    work page doi:10.1016/s0378-7788(00)00114-6 2001

  22. [22]

    Department of Energy, Energyplus building energy simulation pro- gram (2026)

    U.S. Department of Energy, Energyplus building energy simulation pro- gram (2026). URLhttps://energyplus.net/ 17

    work page 2026