Pith. sign in

REVIEW 2 major objections 4 minor 23 references

Large-Load Demand Flexibility as Virtual Storage

T0 review · 2 major / 4 minor · reviewed 2026-07-11 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Large flexible industrial loads can be treated as charge-only virtual storage, so they co-dispatch with batteries in one linear program.

desk verdict Clean, elementary set equivalence that actually lets large loads and BESS share one LP; case-study numbers are optimistic but the math holds under the stated base model. read the letter →

arxiv 2607.04564 v1 pith:S5G5GEPI submitted 2026-07-06 eess.SY cs.SY

Chandan Chaudhary , Mohammed Ben-Idris , Joydeep Mitra This is my paper · ORCID
classification eess.SYcs.SY
keywords demandflexibilityenergystoragelargeloadsvirtualco-dispatchoperationalplanningRTS-GMLC
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Industrial loads such as electrolyzers, data centers, and aluminum potlines are scheduled with per-interval power floors and ceilings plus a horizon energy target, while batteries are scheduled with state-of-charge dynamics. The paper shows these two descriptions are equivalent once load curtailment is rewritten as the charge trajectory of a virtual storage device that never discharges and has unity accounting efficiency in the grid balance. Because the mapping is exact, a portfolio of such loads plus a co-located battery can be optimized together in a single linear program whose load-side constraints scale only with the horizon length, not with the number of loads. Production and service costs sit outside the abstraction and enter only through opportunity costs of curtailment. On the IEEE RTS-GMLC test system the virtual-storage side supplies most of the joint procurement savings, the two resources largely serve different price intervals, and the shadow price of the curtailment budget tracks the start of the peak-price band rather than the daily price spike.

What carries the argument

Virtual Storage Equivalence: the curtailment trajectory of a large flexible load is rewritten as the non-decreasing cumulative charge state of a charge-only virtual storage device, making the two feasibility sets identical and allowing Minkowski-sum aggregation of many loads into an O(T) outer set that co-dispatches with a physical battery.

What would settle it

Re-run the co-dispatch LP on the same RTS-GMLC loads after imposing a rolling multi-hour thermal energy constraint for the potline and a strictly positive curtailment opportunity cost; if the virtual-storage and battery savings cease to be additive or the outer approximation becomes systematically infeasible, the claimed practical value of the equivalence is falsified.

Watch

Extended reading notes

Core claim

Every feasible large-load trajectory under power bounds and a horizon energy window is identical to a feasible charge trajectory of a virtual storage device whose power rating equals the load's flexibility depth, whose capacity equals the maximum allowable curtailment energy, and whose accounting efficiency is unity in the grid power balance. The projection of the virtual-storage feasibility set onto the curtailment coordinates recovers the original load feasibility set exactly, so co-located batteries and flexible loads can be jointly optimized without sequential, resource-specific market processes.

Load-bearing premise

The base model treats a single horizon energy window as the only link across time, so intermediate capacity limits are automatic and technology-specific rolling thermal or recovery constraints are left out; the numerical cases also set the opportunity cost of curtailment to zero.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper shows that the feasible curtailment set of a large flexible load under per-interval power bounds and a single-horizon energy window is identical to the charge-trajectory set of a charge-only virtual storage (VS) device with unity accounting efficiency. From this equivalence it builds a Minkowski-sum aggregate for a portfolio of N loads, reducing load-side constraints from O(NT) to O(T), and embeds the aggregate with a co-located BESS in a single co-dispatch LP whose duals supply a joint value-based price. On the IEEE RTS-GMLC with three representative loads (electrolyzer, data center, potline) and exogenous day-ahead LMPs, co-dispatch savings are dominated by VS under zero curtailment opportunity cost; savings appear additive because the two resources occupy non-overlapping price intervals, and the curtailment-budget shadow price tracks peak-band onset rather than the daily peak.

Significance. If the result holds under the stated base model, the paper supplies a clean, implementable bridge between two previously incompatible scheduling formulations. The set equivalence itself is elementary once monotonic cumulative curtailment is observed, but that is a strength: it is exact, parameter-free within the base abstraction, and immediately yields a joint LP and a constraint-count reduction that scales independently of portfolio size. The RTS-GMLC study is transparent about exogenous prices, zero opportunity cost, and permanent disaggregation infeasibility, so the empirical numbers are interpretable as price-taking upper bounds rather than oversold network-constrained value. The framing is useful for operational planning and for thinking about settlement signals based on the curtailment-budget dual.

major comments (2)
  1. [Section IV, Tables II–III] Section IV and Table II–III: all reported savings and the claim that “VS delivers the dominant share” are obtained with q_t = 0. Under that choice the LP exhausts the full 3,720 MWh budget every day, VS is strictly preferred to BESS at any positive price, and additivity follows mechanically. The paper notes that nonzero opportunity cost would shift contributions, but no sensitivity is provided. Because the abstract and conclusion present dominance and additivity as empirical findings, at least a one-parameter sweep on q (or a simple piecewise-constant opportunity-cost schedule) is needed to show that the qualitative ranking survives realistic production/SLA costs.
  2. [Section III.B, IV.C–D] Section III.B and IV.C–D: the inter-area portfolio is constructed to violate the proportionality condition, and disaggregation is infeasible on all 14 days. Consequently every co-dispatch result is computed on the strict outer set F_outer, which overstates simultaneous deliverable flexibility. The paper correctly flags this as diagnostic of portfolio composition, yet the procurement-cost savings and efficiency-advantage figures are still reported as achieved system value. Either (i) quantify the gap by solving a second-stage disaggregation-constrained LP (or an inner approximation) and report the feasible residual, or (ii) restate the numerical claims explicitly as outer-approximation upper bounds throughout the abstract and Section IV.
minor comments (4)
  1. [Figure 1] Figure 1 panel labels appear duplicated/misaligned in the text (e.g., “(a) Load-Only” appears under both the first and second panels). Clean the caption and panel tags so that Load-Only, BESS-Only, and Co-Dispatch are unambiguously identified.
  2. [Section III.A, Eq. (14)] Equation (14) writes the intermediate capacity bounds 0 ≤ s_t ≤ D̄ for all t; the surrounding prose correctly notes that monotonicity makes them redundant once s_T ≤ D̄. A one-sentence remark that the intermediate inequalities may be dropped from the implemented LP would help implementers.
  3. [Section IV.B] The portfolio LMP is a load-weighted average of three nodal prices. A brief justification that this weighting is consistent with the aggregate power balance used in (16) would remove a small ambiguity for readers who expect a single-bus or multi-bus OPF formulation.
  4. [References] References [13]–[15] are listed as 2026 conference papers; if they are still under review or in press, mark them as such so the citation status is clear.

Circularity Check

1 steps flagged · score 1.0 of 10

VS equivalence is true by construction of the device parameters from the load bounds; no fitted predictions or load-bearing self-citation loops.

  1. self definitional [Section III.A, Virtual Storage Equivalence and F_VS definition (14)]
    "The virtual storage (VS) device is defined by this charge-only trajectory, with initial state s_0 = 0, power rating ΔP, capacity C_VS = D̄, minimum terminal energy D, and unity accounting efficiency. ... Virtual Storage Equivalence: In the base scheduling abstraction, the projection of F_VS onto the deviation coordinates coincides exactly with F_δ. The set of all curtailment trajectories consistent with the base model is identical to the set of charge trajectories of the VS device."

    F_VS is constructed with exactly the cumulative-sum dynamics, power bounds, and terminal energy window taken from F_δ (via (7)–(8) and (13)). The claimed identity therefore holds by the definition of the VS parameters and the elementary monotonicity of s_t; no independent derivation is required or supplied beyond that construction.

full rationale

The central claim (Section III.A) is an exact set identity: the projection of the defined F_VS onto deviation coordinates equals F_δ under the base model of power bounds (1) and single-horizon energy window (2)/(6). This holds because the VS device is parametrized directly from those same bounds (C_VS = D̄, power rating = ΔP, terminal energy in [D, D̄], unity efficiency, charge-only via cumulative s_t) and because cumulative curtailment is nondecreasing, so intermediate capacity bounds are automatic. The paper presents this as a constructive reformulation that enables co-dispatch, not as an independent first-principles prediction or empirical discovery. Aggregation (Minkowski outer set) and the joint LP are standard and do not close a loop. Case-study savings, efficiency advantage, and λ_D̄ values are outputs of the LP under exogenous LMPs and q_t = 0; they are not re-used as inputs. Self-citations [13]–[15] address resource-adequacy context for large loads and are not invoked to justify the equivalence or uniqueness of the VS mapping. No parameters are fitted to data and then re-presented as predictions. The modeling choices (base energy window only, q_t = 0) limit empirical scope but do not make the claimed identity circular. Score remains low because the contribution is an explicit, self-contained reformulation rather than a hidden tautology.

Assumptions & free parameters 4 free parameters · 5 assumptions · 1 invented entities

The central equivalence rests on a small set of modeling axioms that define the base load feasibility set and the accounting convention for curtailment; free parameters are the numerical design-point values chosen for the three load classes and the BESS, plus the zero opportunity-cost assumption used in the numerical study. The only invented entity is the virtual-storage device itself, introduced as a pure mathematical construct.

free parameters (4)
  • VS parameters (ΔP, D, D̄) for electrolyzer, data center, potline = Table I: 180/240/1200, 150/600/1800, 45/360/720 MWh
    Design-point choices ‘within ranges reported’ in the cited sources; they fix the aggregate budget of 3 720 MWh that is exhausted every study day and therefore drive the reported savings magnitudes.
  • curtailment opportunity cost q_t = 0
    Set identically to zero for the entire case study; nonzero q would reorder VS versus BESS dispatch and shrink the claimed VS dominance.
  • BESS rating, efficiency, initial SOC = 200 MW / 400 MWh, η_c=η_d=0.95, e0=200 MWh
    Chosen by the authors (200 MW / 400 MWh, η=0.95, e0=200 MWh); size relative to the VS budget determines the additive-savings observation.
  • 14-day summer study window and portfolio LMP weights = July 5–18 2020; weights 200/250/300 MW
    July 5–18 2020 and load-weighted average of three nodal LMPs; different seasons or weightings could alter the shadow-price stability claim.
assumptions (5)
  • domain assumption Large-load feasibility is completely described by per-interval power bounds p̲ ≤ p_t ≤ p̄ and a single horizon energy window E_min ≤ Σ p_t Δt ≤ E_max (base model).
    Stated in II.A–B; technology-specific rolling thermal or recovery constraints are deferred.
  • domain assumption Verified load reduction of δ_t MW is exactly equivalent to a 1 MW injection in the nodal power balance (unity accounting efficiency).
    Core of the VS definition in III.A; production/SLA costs are declared outside the abstraction.
  • standard math Cumulative curtailment s_t is non-decreasing, so s_T ≤ D̄ automatically implies s_t ≤ D̄ for all intermediate t.
    Used to drop intermediate capacity inequalities in the VS feasibility set (14).
  • domain assumption Operators are price-takers with respect to exogenous day-ahead LMPs; dispatch quantities do not re-clear the market or alter network flows.
    Explicit in IV opening paragraph; savings are therefore price-taking economic value only.
  • standard math When normalized energy-window ratios are identical across loads, the Minkowski sum equals the outer box F_outer; otherwise F_outer is a (possibly strict) outer approximation requiring ex-post disaggregation.
    Aggregation theorem in III.B; the case-study portfolio deliberately violates proportionality.
invented entities (1)
  • Virtual Storage (VS) device
    purpose: Provide a charge-only storage object whose feasible trajectories are identical to the curtailment trajectories of a large flexible load, thereby enabling a common LP with physical BESS.
    Defined in III.A from the load’s own power and energy bounds; it is a pure accounting construct with no physical counterpart and no independent empirical handle beyond the original load constraints.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Large-Load Demand Flexibility as Virtual Storage." pith.science (2026). https://pith.science/paper/S5G5GEPI

@misc{pith2026260704564,
  author       = {Pith},
  title        = {Pith review of: Large-Load Demand Flexibility as Virtual Storage},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S5G5GEPI}},
  note         = {Machine review of arXiv:2607.04564}
}
read the original abstract

Water electrolysis plants, hyperscale data centers, and aluminum potlines represent gigawatts of demand-side flexibility for bulk power system balancing, operational planning, and procurement services. Such loads are scheduled through per-interval power bounds and horizon energy windows, whereas co-located battery energy storage systems (BESS) operate under state-of-charge dynamics. The two formulations share no common mathematical structure, and the joint procurement value of co-located loads and storage goes unrealized as a result. This paper establishes the connection between the two formulations through a virtual storage (VS) equivalence. Every feasible large-load trajectory under power-bound and energy-window constraints is a valid charge trajectory of a VS device that operates at unity accounting efficiency in the grid power balance. Production and service-level costs lie outside this abstraction and enter the dispatch through curtailment opportunity costs. For a portfolio co-located with a BESS, aggregation reduces the constraint count from O(NT) to O(T) and yields a co-dispatch price for both resources. Validation on the IEEE RTS-GMLC with three representative load classes shows that virtual storage delivers the dominant share of joint procurement savings. In the tested case, savings are additive because the two resources dispatch to non-overlapping intervals, and the curtailment shadow price tracks the peak-price band onset rather than the daily peak price.

Figures

Figures reproduced from arXiv: 2607.04564 by the authors.

Figure 1
Figure 1. Net load (MW) and LMP ($/MWh) for three scheduling scenarios, July 5. BESS charge/discharge intervals are shaded in (b) and (c). [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Daily procurement cost savings by scenario, July 5–18. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

23 extracted references

  1. [1]

    Essential reliability services task force – measures framework report,

    North American Electric Reliability Corporation (NERC), “Essential reliability services task force – measures framework report,” NERC, Tech. Rep., 2015

    2015

  2. [2]

    Fast frequency response concepts and bulk power system reliability needs,

    NERC, “Fast frequency response concepts and bulk power system reliability needs,” NERC Inverter-Based Resource Performance Task Force, Tech. Rep., 2020

    2020

  3. [3]

    Analyzing operational flexibility of electric power systems,

    A. Ulbig and G. Andersson, “Analyzing operational flexibility of electric power systems,”International Journal of Electrical Power & Energy Systems, vol. 72, pp. 155–164, 2015

    2015

  4. [4]

    Demand response and smart grids—A survey,

    P. Siano, “Demand response and smart grids—A survey,”Renewable and Sustain- able Energy Reviews, vol. 30, pp. 461–478, 2014

    2014

  5. [5]

    Aggregate flexibility of thermostatically controlled loads,

    H. Hao, B. M. Sanandaji, K. Poolla, and T. L. Vincent, “Aggregate flexibility of thermostatically controlled loads,”IEEE Transactions on Power Systems, vol. 30, no. 1, pp. 189–198, 2014

    2014

  6. [6]

    A geometric approach to aggregate flexibility modeling of thermostatically controlled loads,

    L. Zhao, W. Zhang, H. Hao, and K. Kalsi, “A geometric approach to aggregate flexibility modeling of thermostatically controlled loads,”IEEE Transactions on Power Systems, vol. 32, no. 6, pp. 4721–4731, 2017

    2017

  7. [7]

    Flexibility envelopes for power system operational planning,

    H. Nosair and F. Bouffard, “Flexibility envelopes for power system operational planning,”IEEE Transactions on Sustainable Energy, vol. 6, no. 3, pp. 800–809, 2015

    2015

  8. [8]

    On the ability of PEM water electrolysers to provide power grid services,

    L. Allidières, A. Brisse, P. Millet, S. Valentin, and M. Zeller, “On the ability of PEM water electrolysers to provide power grid services,”International Journal of Hydrogen Energy, vol. 44, no. 20, pp. 9690–9700, 2019

    2019

Show all 23 references
  1. [9]

    Grid balancing with a large-scale electrolyser providing primary reserve,

    A. E. Samani, A. D’Amicis, J. D. M. De Kooning, D. Bozalakov, P. Silva, and L. Vandevelde, “Grid balancing with a large-scale electrolyser providing primary reserve,”IET Renewable Power Generation, vol. 14, no. 16, pp. 3070–3078, 2020

    2020

  2. [10]

    Opportunities and challenges for data center demand response,

    A. Wierman, Z. Liu, I. Liu, and H. Mohsenian-Rad, “Opportunities and challenges for data center demand response,” inProc. International Green Computing Con- ference (IGCC), 2014, pp. 1–10

    2014

  3. [11]

    Data center demand response: Avoiding the coincident peak via workload shifting and local generation,

    Z. Liu, A. Wierman, Y . Chen, B. Razon, and N. Chen, “Data center demand response: Avoiding the coincident peak via workload shifting and local generation,” Performance Evaluation, vol. 70, no. 10, pp. 770–791, 2013

    2013

  4. [12]

    Bidding strategy in energy and spinning reserve markets for aluminum smelters’ demand response,

    X. Zhang and G. Hug, “Bidding strategy in energy and spinning reserve markets for aluminum smelters’ demand response,” inProc. IEEE Power & Energy Society Innovative Smart Grid Technologies (ISGT), 2015, pp. 1–5

    2015

  5. [13]

    Resource adequacy risk in correlated large loads,

    C. Chaudhary, A. Abdelkader, M. Ben-Idris, and J. Mitra, “Resource adequacy risk in correlated large loads,” in2026 19th International Conference on Probabilistic Methods Applied to Power Systems (PMAPS), Salt Lake City, UT, USA, Sep. 2026

    2026

  6. [14]

    Spatial load correlation in AI data-center-dominated power systems,

    C. Chaudhary, A. Abdelkader, Y . Pei, M. Benidris, and J. Mitra, “Spatial load correlation in AI data-center-dominated power systems,” in2026 IEEE Power & Energy Society General Meeting (PES GM), Montréal, QC, Canada, Jul. 2026

    2026

  7. [15]

    Modal analysis of spatial load correlation in AI data center-dominated power systems,

    C. Chaudhary, M. Murillo, M. Ben-Idris, J. Mitra, D. Pandit, and A. Bera, “Modal analysis of spatial load correlation in AI data center-dominated power systems,” in2026 IEEE 8th International Conference on Smart Energy Systems and Technologies (SEST), Ciudad Real, Spain, Sep. 2026

    2026

  8. [16]

    Large-scale integration of deferrable demand and renewable energy sources,

    A. Papavasiliou and S. S. Oren, “Large-scale integration of deferrable demand and renewable energy sources,”IEEE Transactions on Power Systems, vol. 29, no. 1, pp. 489–499, 2014

    2014

  9. [17]

    Using a transactive energy framework: Providing grid services from smart buildings,

    F. Rahimi and A. Ipakchi, “Using a transactive energy framework: Providing grid services from smart buildings,”IEEE Electrification Magazine, vol. 4, no. 4, pp. 23–29, 2016

    2016

  10. [18]

    A society of devices: Integrating intelligent distributed resources with transactive energy,

    K. Kok and S. Widergren, “A society of devices: Integrating intelligent distributed resources with transactive energy,”IEEE Power and Energy Magazine, vol. 14, no. 3, pp. 34–45, 2016

    2016

  11. [19]

    Strategic capacity withholding by energy storage in electricity markets,

    Y . Ye, D. Papadaskalopoulos, R. Moreira, and G. Strbac, “Strategic capacity withholding by energy storage in electricity markets,” in2017 IEEE Manchester PowerTech, 2017, pp. 1–6

    2017

  12. [20]

    The IEEE reliability test system: A proposed 2019 update,

    C. Barrows, A. Bloom, A. Ehlen, J. Ikäheimo, J. Jorgenson, D. Krishnamurthy, J. Lau, B. McBennett, M. O’Connell, E. Preston, A. Staid, G. Stephen, and J.- P. Watson, “The IEEE reliability test system: A proposed 2019 update,”IEEE Transactions on Power Systems, vol. 34, no. 3, ...

    2019

  13. [21]

    Matute, J

    G. Matute, J. M. Yusta, and L. C. Correas, “Techno-economic modelling of water electrolysers in the range of several MW to provide grid services while generating hydrogen for different applications: A case study in Spain applied to mobility with FCEVs,”International Journal of...

    2019

  14. [22]

    Parametric modelling and optimization of alkaline electrolyzer for the production of green hydrogen,

    S. Niroula, C. Chaudhary, A. Subedi, and B. S. Thapa, “Parametric modelling and optimization of alkaline electrolyzer for the production of green hydrogen,”IOP Conference Series: Materials Science and Engineering, vol. 1279, p. 012005, 2023

    2023

  15. [23]

    Parallelizing the dual revised simplex method,

    Q. Huangfu and J. A. J. Hall, “Parallelizing the dual revised simplex method,” Mathematical Programming Computation, vol. 10, pp. 119–142, 2018

    2018

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.