pith. sign in

arxiv: 2605.26392 · v1 · pith:IGNLNN3Rnew · submitted 2026-05-25 · 🧮 math.OC

Integrating Hydrogen into Ontario's Energy Hub: A Robust, Carbon-Aware Framework for Power-Heat-Transport

Hossein Mirzaee , Mostafa Mostafavi Sani , Hamed Samarghandi This is my paper

Pith reviewed 2026-06-29 20:06 UTC · model grok-4.3

classification 🧮 math.OC
keywords energy hubhydrogen integrationrobust optimizationcarbon-aware planningMILPdecarbonizationOntariouncertainty
0
0 comments X

The pith

A robust MILP framework for Ontario's multi-energy hub projects electrolyzer capacity rising from 300 MW to 3,800 MW and storage to 37,000 MWh by 2050 under carbon policies.

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

The paper develops a mixed-integer linear program that co-optimizes electricity, heating, cooling, and transport services inside a grid-connected hub that includes a dedicated hydrogen sub-system. The model runs over a 25-year horizon and is extended with budgeted robust optimization to protect against uncertainty in operations. When solved for Ontario under both a carbon tax and a net-zero constraint, it produces large long-term growth in electrolyzer and hydrogen storage capacity. The robust version keeps all plans feasible but raises total system cost by 6.6 to 9 percent while slightly lowering renewable share and hydrogen use. This demonstrates how uncertainty can be hedged inside integrated decarbonization planning without eliminating the hydrogen pathway.

Core claim

The developed scheme applied to Ontario indicates substantial long-term hydrogen expansion, with the electrolyzer capacity increasing from 300 MW (2025) to 3,800 MW (2050) and hydrogen storage from 2,000 MWh to 37,000 MWh (2050), accompanied by sharply higher hydrogen production. Robust solutions preserve feasibility but incur a 6.6-9.0% robustness premium in total cost.

What carries the argument

A 25-year MILP for a grid-connected energy hub extended by budgeted robust optimization; the hydrogen sub-hub (electrolyzer plus storage) supplies transport demand and provides operational flexibility across electricity, heat, cooling, and transport services.

Load-bearing premise

The 25-year projections rest on assumed trajectories for energy service demands, technology capital and operating costs, renewable resource availability, and conversion efficiencies across the 2025-2050 horizon.

What would settle it

Actual Ontario electrolyzer installations by 2035 that fall well below the model's projected path under comparable carbon-tax or net-zero policies would falsify the expansion trajectory.

Figures

Figures reproduced from arXiv: 2605.26392 by Hamed Samarghandi, Hossein Mirzaee, Mostafa Mostafavi Sani.

Figure 1
Figure 1. Figure 1: Carrier-flow schematic of the proposed multi-energy hub. Solid rectangles denote sources and conversion technologies, dashed nodes denote storage units, shaded rectangles denote end-use demands, and colored arrows indicate electricity, heat, cooling, and hydrogen flows. losses captured through efficiency parameters. EV and FCEV demands are represented in an aggregate form, without considering individual mo… view at source ↗
Figure 2
Figure 2. Figure 2: Energy supply by source during the representative days 21 [PITH_FULL_IMAGE:figures/full_fig_p021_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Energy supply composition over representative cold and warm days at the beginning (2025) and end (2050) of the 25-year planning horizon, shown separately for the carbon tax scenario (panels a–b) and the net zero scenario (panels c–d). 4.1.2 Long-term hydrogen build-out (2025–2050) [PITH_FULL_IMAGE:figures/full_fig_p022_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Transition under the carbon tax pathway: fossil fuel energy use decrease while both electricity and hydrogen supplies increase over 2025–2050 4.1.3 Deterministic policy effect: carbon tax vs net-zero [PITH_FULL_IMAGE:figures/full_fig_p023_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Schematic long-run trajectories (2025–2050) illustrating the expected directional effects of carbon tax versus net zero policies: (top) emissions decline under both policies, with net zero enforcing a steeper reduction toward near-zero by 2050; (middle) fossil fuel use declines in both cases, with net zero requiring deeper phase-down by 2050; (bottom) system cost, excluding carbon costs, increases more und… view at source ↗
Figure 6
Figure 6. Figure 6: Sensitivity tornado: one-at-a-time (OAT) perturbations showing the relative influence of candidate parameters on total cost. Bars indicate the range of cost changes under the low and high perturbation levels. 26 [PITH_FULL_IMAGE:figures/full_fig_p026_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Robustness premium as the uncertainty budget Γ increases (2025 baseline). Larger Γ implies protection against more simultaneous adverse deviations, increasing the objective value. 4.3.3 Deterministic versus robust comparison and managerial interpretation The comparison between deterministic and robust results provides quantification of the robust￾ness premium and helps identify the mechanisms through which… view at source ↗
Figure 8
Figure 8. Figure 8: Carbon tax case (2025 baseline): deterministic vs robust total cost response under one-at-a-time perturbations in (a) total market demand and (b) hydrogen energy price. (a) Total market demand (Det vs Rob). (b) Hydrogen energy price (Det vs Rob) [PITH_FULL_IMAGE:figures/full_fig_p032_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Net zero case (2025 baseline): deterministic vs robust total cost response under one-at-a-time perturbations in (a) total market demand and (b) hydrogen energy price. Figures 8 and 9 illustrate the variation in total costs when the total market demand changes. In all cases, it can be seen that the robust cost curves remain above the deterministic curves over the range of perturbations. These plots illustra… view at source ↗
Figure 10
Figure 10. Figure 10: Total cost trajectories (2025–2050): carbon tax policy (top) and net zero policy (bottom), comparing deterministic and robust solutions [PITH_FULL_IMAGE:figures/full_fig_p033_10.png] view at source ↗
read the original abstract

Decarbonizing electricity generation, heating, and transportation simultaneously requires integrated planning tools that can coordinate multiple energy production sources and demand points while remaining reliable under uncertainty. This paper develops a carbon-aware and uncertainty-resilient optimization framework for a grid-connected multiple source hub that co-optimizes electricity, heating, cooling, and transport energy services with an explicit hydrogen sub-hub. The proposed model is formulated as an MILP over a 25-year planning horizon (2025-2050). The hub integrates renewable electricity (Photovoltaic and wind), dispatchable resources (including natural-gas-based conversion), storage systems, demand response, and a hydrogen subsystem comprising an electrolyzer and hydrogen storage to supply hydrogen-vehicle demand and provide temporal flexibility. Two policy archetypes are examined: a Carbon Tax (price instrument), and a Net-Zero pathway (quantity instrument). To hedge feasibility-critical operational uncertainty, the deterministic model is extended using budgeted robust optimization and a tunable uncertainty budget. The developed scheme is applied to the province of Ontario, Canada; the results indicate substantial long-term hydrogen expansion, with the electrolyzer capacity increasing from 300~MW (2025) to 3,800~MW (2050) and hydrogen storage from 2,000~MWh to 37,000~MWh (2050), accompanied by sharply higher hydrogen production. Compared with deterministic solutions, robust solutions preserve feasibility under adverse realizations but incur a moderate robustness premium of approximately 6.6-9.0\% in total cost across the policy cases studied, while slightly reducing hydrogen utilization and renewable share and increasing reliance on dispatchable balancing.

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 / 0 minor

Summary. The paper develops a carbon-aware MILP for co-optimizing power, heat, cooling, and transport services in a grid-connected Ontario energy hub that includes a hydrogen sub-hub (electrolyzer and storage) over a 2025-2050 horizon. It examines two policy cases (carbon tax and net-zero), extends the model with budgeted robust optimization to hedge operational uncertainty, and reports specific results for Ontario: electrolyzer capacity rising from 300 MW to 3,800 MW, hydrogen storage from 2,000 MWh to 37,000 MWh, sharply higher hydrogen production, and a 6.6-9.0% robustness premium in total cost relative to deterministic solutions.

Significance. If the exogenous input trajectories and model validation hold, the work supplies a concrete, multi-sector planning tool that couples hydrogen with robustness under carbon policies; the budgeted-robust extension and Ontario-scale numeric outputs are the primary contributions.

major comments (2)
  1. [Abstract] Abstract and results: the headline capacity expansions (electrolyzer 300 MW → 3,800 MW; storage 2,000 MWh → 37,000 MWh) and 6.6-9.0% robustness premium are direct outputs of the budgeted robust MILP; these quantities are fully determined by the assumed 25-year time series for energy-service demands, technology costs/efficiencies, renewable availability, and carbon parameters, yet the manuscript provides neither sourcing nor sensitivity analysis on these exogenous inputs.
  2. [Model Formulation] Model and data sections: the abstract states that an MILP is solved under two policies plus robust extension and then reports concrete capacity and cost numbers, but without the explicit formulation, data sources, or validation against historical Ontario loads, it is impossible to confirm that the model supports the reported numeric outcomes.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. The comments highlight important aspects of transparency that we will address in revision. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract and results: the headline capacity expansions (electrolyzer 300 MW → 3,800 MW; storage 2,000 MWh → 37,000 MWh) and 6.6-9.0% robustness premium are direct outputs of the budgeted robust MILP; these quantities are fully determined by the assumed 25-year time series for energy-service demands, technology costs/efficiencies, renewable availability, and carbon parameters, yet the manuscript provides neither sourcing nor sensitivity analysis on these exogenous inputs.

    Authors: We agree that explicit sourcing and sensitivity analysis on the exogenous inputs would improve reproducibility. The Case Study section describes the input trajectories at a high level, but we will add a dedicated data appendix with full references (e.g., IESO long-term energy plans for Ontario demands, IEA and NREL cost projections, and Environment and Climate Change Canada carbon parameters). We will also include sensitivity results on demand growth rates, renewable capacity factors, and carbon price levels to show how the reported electrolyzer and storage capacities respond. These changes will be incorporated in the revised manuscript. revision: yes

  2. Referee: [Model Formulation] Model and data sections: the abstract states that an MILP is solved under two policies plus robust extension and then reports concrete capacity and cost numbers, but without the explicit formulation, data sources, or validation against historical Ontario loads, it is impossible to confirm that the model supports the reported numeric outcomes.

    Authors: The Model Formulation section presents the deterministic MILP and its budgeted-robust extension, including the objective, power/heat/transport balance constraints, and hydrogen sub-hub equations. However, to address the concern directly, we will expand this section with the complete mathematical formulation (all variables, constraints, and the robust counterpart) placed in the main text or a supplementary appendix. We will also add a model validation subsection that compares 2025-period outputs against historical Ontario electricity and natural-gas load data from public IESO and Statistics Canada sources. Data tables with units and sources will be added to the Case Study section. These revisions will be made without changing the core numeric results. revision: yes

Circularity Check

0 steps flagged

No circularity: MILP outputs are direct solutions to exogenous input trajectories

full rationale

The paper presents a standard MILP capacity-expansion model over a 25-year horizon, extended with budgeted robust optimization to handle uncertainty. Reported results (electrolyzer capacity from 300 MW to 3,800 MW, storage from 2,000 MWh to 37,000 MWh, 6.6-9.0% robustness premium) are explicit solutions of the optimization given assumed exogenous time series for demands, costs, renewable availability, and efficiencies. No equation reduces any output to a quantity defined by the same output, no fitted parameter is relabeled as a prediction, and no load-bearing premise rests on self-citation or imported uniqueness. The derivation chain is therefore self-contained and non-circular.

Axiom & Free-Parameter Ledger

4 free parameters · 2 axioms · 0 invented entities

Based solely on the abstract, the central numerical claims rest on a large set of exogenous input time series (demands, costs, efficiencies) that are not enumerated; the model itself introduces no new physical entities.

free parameters (4)
  • technology cost and efficiency trajectories
    Capital and operating costs plus conversion efficiencies for PV, wind, electrolyzers, storage, and gas plants over 25 years are required inputs to the MILP.
  • energy service demand forecasts
    Projected electricity, heat, cooling, and hydrogen-vehicle demands for each year are required to size the hub.
  • carbon policy parameters
    Tax rates or net-zero emission caps are varied across the two policy archetypes.
  • uncertainty budget parameter
    The tunable budget that controls how much deviation from nominal values is protected against in the robust formulation.
axioms (2)
  • domain assumption The multi-carrier energy system can be represented as a single hub with linear conversion and storage balances
    Standard modeling choice for energy-hub MILPs; invoked to allow co-optimization of power, heat, cooling, and hydrogen.
  • domain assumption Operational uncertainty can be captured by a budgeted polyhedral uncertainty set
    Core assumption of the robust optimization extension; allows tractable MILP reformulation.

pith-pipeline@v0.9.1-grok · 5840 in / 1759 out tokens · 45811 ms · 2026-06-29T20:06:03.843696+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

13 extracted references · 12 canonical work pages

  1. [1]

    1016/j.eneco.2026.109221

    Will hydrogen prolong the price coupling between gas and electricity? Energy Economics , 109221doi:10. 1016/j.eneco.2026.109221. Dilek, G., Teixido, J.J., Serrano, M.,

    work page arXiv 2026

  2. [2]

    Energy Economics , 109174doi:10.1016/j.eneco.2026

    Carbon pricing effects on renewables: Evidence from california’s electricity market. Energy Economics , 109174doi:10.1016/j.eneco.2026. 109174. Gabrielli, P., Poluzzi, A., Kramer, G.J., Spiers, C., Mazzotti, M., Gazzani, M., 2020a. Seasonal 35 energy storage for zero-emissions multi-energy systems via underground hydrogen storage. Renewable and Sustainabl...

    work page doi:10.1016/j.eneco.2026 2026

  3. [3]

    URL:https://www.canada.ca/en/department-finance/ news/2025/03/removing-the-consumer-carbon-price-effective-april-1-2025.html

    Department of Finance Canada. URL:https://www.canada.ca/en/department-finance/ news/2025/03/removing-the-consumer-carbon-price-effective-april-1-2025.html. accessed: May

    2025

  4. [4]

    Energy & Environmental Science 13, 1909–1922

    The value of seasonal energy storage technologies for the integration of wind and solar power. Energy & Environmental Science 13, 1909–1922. doi:10.1039/D0EE00771D. Hou, S., Fan, Y., Yi, B.W.,

    work page doi:10.1039/d0ee00771d 1909

  5. [5]

    Ioannou, A., Fuzuli, G., Brennan, F., Yudha, S.W., Angus, A.,

    doi:10.3390/mining5010009. Ioannou, A., Fuzuli, G., Brennan, F., Yudha, S.W., Angus, A.,

    work page doi:10.3390/mining5010009

  6. [6]

    Energy Economics , 109226doi:10.1016/j.eneco.2026.109226

    How to sustainably spur r&d of energy storage? policy options and combinations. Energy Economics , 109226doi:10.1016/j.eneco.2026.109226. Khakimov, R., Moskvin, A., Zhdaneev, O.,

    work page doi:10.1016/j.eneco.2026.109226 2026

  7. [7]

    arXiv preprint arXiv:2602.02333

    A two-stage stochastic optimization model for the equitable deployment of fixed and mobile electric vehicle charging stations. arXiv preprint arXiv:2602.02333 . Nunes, J.B., Mahmoudi, N., Saha, T.K., Chattopadhyay, D.,

    work page arXiv

  8. [8]

    Energy Economics , 108981doi:10.1016/j.eneco.2025.108981

    Hedging hydrogen: Plan- ning and contracting under uncertainty for a green hydrogen producer. Energy Economics , 108981doi:10.1016/j.eneco.2025.108981. Papadimitriou, C., Di Somma, M., Charalambous, C., Caliano, M., Palladino, V., Cortes Borray, A., Gonz´ alez-Garrido, A., Ruiz, N., Graditi, G., 2023a. A comprehensive review of the design 37 and operation...

    work page doi:10.1016/j.eneco.2025.108981 2025

  9. [9]

    Power-to-hydrogen as seasonal energy storage: An uncertainty analysis for optimal design of low-carbon multi-energy systems

    Petkov, I., Gabrielli, P., 2020a. Power-to-hydrogen as seasonal energy storage: An uncertainty analysis for optimal design of low-carbon multi-energy systems. Applied Energy 274, 115197. doi:10.1016/j.apenergy.2020.115197. Petkov, I., Gabrielli, P., 2020b. Power-to-hydrogen as seasonal energy storage: An uncertainty analysis for optimal design of low-carb...

    work page doi:10.1016/j.apenergy.2020.115197 2020

  10. [10]

    Energy 317, 134678

    Modeling a realistic integrated energy hub with growing demand for electric vehicles: The case of the province of Ontario, Canada. Energy 317, 134678. doi:10.1016/j.energy.2025.134678. 38 Staffell, I., Scamman, D., Abad, A.V., Balcombe, P., Dodds, P.E., Ekins, P., Shah, N., Ward, K.R.,

    work page doi:10.1016/j.energy.2025.134678 2025

  11. [11]

    IEEE Access 8, 2395–2408

    Coordinated planning with predetermined renew- able energy generation targets using extended two-stage robust optimization. IEEE Access 8, 2395–2408. Tsao, Y.C., Vu, T.L., 2023a. Distributed energy storage system planning in relation to renewable energy investment. Renewable Energy 218, 119271. Tsao, Y.C., Vu, T.L., 2023b. Distributed energy storage syste...

    work page doi:10.1016/j.renene.2023.119271 2023

  12. [12]

    Applied Energy 302, 117579

    Unlocking emerging impacts of carbon tax on integrated energy systems through supply and demand co-optimization. Applied Energy 302, 117579. doi:10.1016/j.apenergy.2021.117579. Wu, Q., Chen, M., Ren, H., Li, Q., Gao, W.,

    work page doi:10.1016/j.apenergy.2021.117579 2021

  13. [13]

    Renewable Energy 227, 120489

    Collaborative modeling and optimization of energy hubs and multi-energy network considering hydrogen energy. Renewable Energy 227, 120489. doi:10.1016/j.renene.2024.120489. Yu, H., Niu, S., Shang, Y., Shao, Z., Jia, Y., Jian, L.,

    work page doi:10.1016/j.renene.2024.120489 2024