pith. sign in

arxiv: 2510.19781 · v2 · submitted 2025-10-22 · 🧮 math.OC · cs.SY· eess.SY

Nodal Capacity Expansion Planning with Flexible Large-Scale Load Siting

Tomas Valencia Zuluaga , Simon Pang , Jean-Paul Watson This is my paper

Pith reviewed 2026-05-18 04:34 UTC · model grok-4.3

classification 🧮 math.OC cs.SYeess.SY
keywords capacity expansion planningstochastic optimizationload sitingprogressive hedgingpower systemsreliability constraintsmixed-integer programmingdatacenters
0
0 comments X

The pith

Incorporating large load siting as reliability tranches into stochastic capacity expansion allows co-optimization of generation, transmission and storage.

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

The paper develops a stochastic nodal model for power system capacity expansion that treats decisions on where to site large new loads as part of the same optimization that chooses new generation, transmission lines and storage. These loads are split into tranches that accept different levels of service interruptions, and the model enforces each tranche's requirements through a constraint on average energy delivered across all operational scenarios. The resulting two-stage mixed-integer program is solved with an augmented progressive hedging algorithm that runs in parallel on high-performance computers to keep the computation tractable. Tests on realistic networks assigned to San Diego and South Carolina show how this joint planning changes total system cost and reliability outcomes when the new loads are data centers or direct air capture plants. A sympathetic reader cares because electricity demand from very large facilities is growing quickly and traditional planning that fixes load locations in advance may produce unnecessarily expensive or brittle grids.

Core claim

We propose explicitly incorporating large-scale load siting into a stochastic nodal power system capacity expansion planning model that concurrently co-optimizes generation, transmission and storage expansion. The potential operational flexibility of some of these large loads is also taken into account by considering them as consisting of a set of tranches with different reliability requirements, which are modeled as a constraint on expected served energy across operational scenarios. We implement our model as a two-stage stochastic mixed-integer optimization problem with cross-scenario expectation constraints and solve it via an augmented Progressive Hedging Algorithm on a high-performance,

What carries the argument

Tranche representation of large loads together with expected served energy constraints across operational scenarios, which embeds operational flexibility directly into the investment decisions.

If this is right

  • Joint optimization of load siting with infrastructure choices produces plans that better align new demand with available supply and transmission capacity.
  • Tranches with different reliability targets can be served without requiring uniform high reliability for every megawatt of new load.
  • The value of proactive load siting appears in lower total system costs and improved reliability metrics on the tested geographic cases.

Where Pith is reading between the lines

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

  • The same tranche-and-expectation structure could be applied to other large flexible demands whose locations are still under negotiation.
  • Regulators may need to coordinate siting approvals for large loads with the timing of grid planning studies that use this type of model.
  • Extending the approach to include uncertainty in the very locations of future loads would be a direct next modeling step.

Load-bearing premise

Large loads can be represented accurately enough as discrete tranches with fixed reliability requirements whose expected served energy constraints capture their operational flexibility, and the augmented progressive hedging solver produces usable solutions on the instances tested even without proven convergence.

What would settle it

Running the model on the San Diego or South Carolina test cases and finding that the resulting total system cost is no lower and reliability metrics are no better than those obtained when load locations are fixed in advance would show the added value does not materialize.

Figures

Figures reproduced from arXiv: 2510.19781 by Jean-Paul Watson, Simon Pang, Tomas Valencia Zuluaga.

Figure 1
Figure 1. Figure 1: Approach used to represent flexibility of large loads. Loads are [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Left: Resource buildout for each test case. Case IIa has by construction the same buildout as Ic. Case IIb has the same buildout plus one natural gas unit co-sited with the datacenter; both are omitted from this figure. Transmission capacity is obtained by summing the capacities of all selected candidate lines or transformers. Right: Total cost and CO2 emissions for each test case.        … view at source ↗
Figure 3
Figure 3. Figure 3: Left: Costs and emissions for cases IIa, IIb, IIc. Note that costs are in a log scale. Right: Achieved and required reliability for each tranche of each large load site built. Different shades of the same color show different tranches of the same site. on the different cases tested. Final costs are normalized with respect to the solution obtained with the EF formulation. For the 24-bus test case, a very go… view at source ↗
read the original abstract

We propose explicitly incorporating large-scale load siting into a stochastic nodal power system capacity expansion planning model that concurrently co-optimizes generation, transmission and storage expansion. The potential operational flexibility of some of these large loads is also taken into account by considering them as consisting of a set of tranches with different reliability requirements, which are modeled as a constraint on expected served energy across operational scenarios. We implement our model as a two-stage stochastic mixed-integer optimization problem with cross-scenario expectation constraints. To overcome the challenge of scalability, we build upon existing work to implement this model on a high performance computing platform and exploit scenario parallelization using an augmented Progressive Hedging Algorithm. The algorithm is implemented using the bounding features of mpisppy, which have shown to provide satisfactory provable optimality gaps despite the absence of theoretical guarantees of convergence. We test our approach to assess the value of this proactive planning framework on total system cost and reliability metrics using realistic testcases geographically assigned to San Diego and South Carolina, with datacenter and direct air capture facilities as large loads.

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

3 major / 2 minor

Summary. The paper proposes a two-stage stochastic MIP formulation for nodal capacity expansion planning that explicitly includes siting decisions for large-scale flexible loads (datacenters and direct air capture facilities). Loads are represented as discrete tranches with distinct reliability requirements enforced via cross-scenario expected served energy constraints; the model jointly optimizes generation, transmission, and storage investments under uncertainty. Scalability is addressed by an augmented Progressive Hedging algorithm implemented in mpisppy, with numerical tests on geographically realistic San Diego and South Carolina instances used to quantify impacts on total system cost and reliability metrics.

Significance. If the numerical evidence is robust, the work would provide a concrete demonstration of the value of proactive, co-optimized load siting within stochastic nodal planning, potentially informing how system operators can leverage operational flexibility from emerging large loads to reduce infrastructure costs while meeting reliability targets. The HPC implementation and use of bounding features in mpisppy also illustrate a practical route for solving large-scale stochastic MIPs with expectation coupling.

major comments (3)
  1. [Numerical Experiments / Algorithm Description] The central cost and reliability claims rest on the quality of the solutions obtained by the augmented PHA. The manuscript asserts that the bounding features yield 'satisfactory provable optimality gaps' on the tested instances, yet no numerical gap values, iteration counts, or bound tightness metrics are reported for the San Diego or South Carolina cases. Without these quantities it is impossible to determine whether the reported deltas are larger than the remaining optimality gap.
  2. [Model Formulation] The modeling choice to represent load flexibility solely through fixed reliability levels assigned to tranches and an expected served energy constraint is load-bearing for the claimed operational flexibility benefit. The paper does not provide sensitivity results on these reliability parameters or compare against alternative flexibility representations (e.g., explicit ramping or curtailment costs), leaving open whether the reported advantages are robust to reasonable variations in tranche definitions.
  3. [Numerical Experiments] The manuscript states that the approach is tested against 'baselines,' but the exact definition of those baselines (e.g., whether they fix load locations a priori, omit tranche flexibility, or use deterministic equivalents) is not specified in sufficient detail to allow replication or to isolate the incremental value of joint siting optimization.
minor comments (2)
  1. [Model Formulation] Clarify the precise mathematical form of the expected served energy constraint (including how non-anticipativity interacts with the cross-scenario expectation) and ensure all symbols are defined before first use.
  2. [Numerical Experiments] Add a short table summarizing instance sizes, number of scenarios, and wall-clock times to give readers a concrete sense of computational scale.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major comment below and indicate the changes we will make in the revised version.

read point-by-point responses

  1. Referee: [Numerical Experiments / Algorithm Description] The central cost and reliability claims rest on the quality of the solutions obtained by the augmented PHA. The manuscript asserts that the bounding features yield 'satisfactory provable optimality gaps' on the tested instances, yet no numerical gap values, iteration counts, or bound tightness metrics are reported for the San Diego or South Carolina cases. Without these quantities it is impossible to determine whether the reported deltas are larger than the remaining optimality gap.

    Authors: We agree that the specific numerical values for optimality gaps, iteration counts, and bound tightness are needed to allow readers to evaluate solution quality. The current manuscript only states that the mpisppy bounding features yield satisfactory gaps without reporting the actual numbers for the two instances. In the revision we will add a dedicated subsection (or table) presenting these metrics for both the San Diego and South Carolina cases, together with a short discussion of how the final gaps compare to the reported cost and reliability differences. revision: yes

  2. Referee: [Model Formulation] The modeling choice to represent load flexibility solely through fixed reliability levels assigned to tranches and an expected served energy constraint is load-bearing for the claimed operational flexibility benefit. The paper does not provide sensitivity results on these reliability parameters or compare against alternative flexibility representations (e.g., explicit ramping or curtailment costs), leaving open whether the reported advantages are robust to reasonable variations in tranche definitions.

    Authors: We acknowledge that the absence of sensitivity analysis on the reliability parameters limits the ability to assess robustness. The tranche reliability levels were chosen to reflect typical requirements for the two load types, but we did not vary them or benchmark against explicit ramping/curtailment cost formulations. In the revised manuscript we will add a sensitivity study on the reliability thresholds and include a brief discussion of the modeling rationale relative to alternative flexibility representations. revision: yes

  3. Referee: [Numerical Experiments] The manuscript states that the approach is tested against 'baselines,' but the exact definition of those baselines (e.g., whether they fix load locations a priori, omit tranche flexibility, or use deterministic equivalents) is not specified in sufficient detail to allow replication or to isolate the incremental value of joint siting optimization.

    Authors: We agree that the baseline definitions require more explicit description to support replication and to isolate the contribution of joint siting. The current text refers to baselines without detailing whether load locations are pre-fixed, whether tranche flexibility is disabled, or how they relate to deterministic runs. In the revision we will expand the numerical-experiments section to define each baseline precisely and to clarify how the incremental value of co-optimized siting is measured. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper defines a two-stage stochastic MIP that explicitly adds large-scale load siting decisions and tranche-based expected-served-energy constraints to a standard co-optimization of generation, transmission, and storage. These elements are introduced as modeling choices rather than derived from or fitted to any output quantity inside the paper. The solution method augments an existing Progressive Hedging implementation in mpisppy; the reported optimality gaps are empirical observations on the test instances and are not used to define or force any of the claimed cost or reliability improvements. No self-definitional, fitted-input, or self-citation-load-bearing reductions appear in the formulation or numerical claims.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard power-system network assumptions and the modeling choice that load flexibility is captured by reliability tranches; no new physical entities are postulated.

free parameters (1)
  • Reliability levels assigned to load tranches
    Chosen values that define how much energy each tranche must receive on average across scenarios.
axioms (1)
  • domain assumption Power system operations can be represented by a finite set of scenarios with nodal balance and transmission constraints.
    Invoked when formulating the two-stage stochastic program with cross-scenario expectation constraints.

pith-pipeline@v0.9.0 · 5718 in / 1301 out tokens · 55751 ms · 2026-05-18T04:34:38.328238+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. To Defer or To Shift? The Role of AI Data Center Flexibility on Grid Interconnection

    eess.SY 2026-04 unverdicted novelty 5.0

    AI data center temporal and spatial flexibility reduces grid investment and operational costs by 3-21% in some locations and load conditions but does not consistently lower required generation capacity and shows dimin...

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages · cited by 1 Pith paper

  1. [1]

    A. F. Johnson and K. Kornecki, Eds.,Implications of Artificial Intelli- gence–Related Data Center Electricity Use and Emissions: Proceedings of a Workshop. Washington, D.C.: National Academies Press, Jul. 2025

    work page 2025

  2. [2]

    Data Center Demand Pipeline Swells to 10 Gigawatts with Potential to Unlock Billions in Benefits for California,

    PG&E, “Data Center Demand Pipeline Swells to 10 Gigawatts with Potential to Unlock Billions in Benefits for California,” Jul. 2025. [Online]. Available: https://investor.pgecorp.com/news-events/press- releases/press-release-details/2025/PGE-Data-Center-Demand-Pipeline- Swells-to-10-Gigawatts-with-Potential-to-Unlock-Billions-in-Benefits- for-California/de...

    work page 2025

  3. [3]

    Impact of direct air capture process flexibility and response to ambient conditions in net-zero transition of the power grid,

    E. O. Arwa and K. R. Schell, “Impact of direct air capture process flexibility and response to ambient conditions in net-zero transition of the power grid,”Applied Energy, vol. 386, p. 125549, May 2025

    work page 2025

  4. [4]

    Analyzing Demand Response in a Dynamic Capacity Expansion Model for the European Power Market,

    H. Mara ˜n´on-Ledesma and A. Tomasgard, “Analyzing Demand Response in a Dynamic Capacity Expansion Model for the European Power Market,”Energies, vol. 12, no. 15, p. 2976, Jan. 2019, publisher: Multidisciplinary Digital Publishing Institute

    work page 2019

  5. [5]

    Capacity Expansion Planning under Uncertainty subject to Expected Energy Not Served Constraints,

    M. Zampara, D. ´Avila, and A. Papavasiliou, “Capacity Expansion Planning under Uncertainty subject to Expected Energy Not Served Constraints,” Jan. 2025, arXiv:2501.17484 [eess]. [Online]. Available: http://arxiv.org/abs/2501.17484

    work page arXiv 2025

  6. [6]

    Quantifying the impact of energy system model resolution on siting, cost, reliability, and emissions for electricity generation,

    A. F. Jacobson, D. L. Mauzerall, and J. D. Jenkins, “Quantifying the impact of energy system model resolution on siting, cost, reliability, and emissions for electricity generation,”Environmental Research: Energy, vol. 1, no. 3, p. 035009, Sep. 2024, publisher: IOP Publishing

    work page 2024

  7. [7]

    Paral- lel computing for power system climate resiliency: Solving a large-scale stochastic capacity expansion problem with mpi-sppy,

    T. Valencia Zuluaga, A. Musselman, J.-P. Watson, and S. S. Oren, “Paral- lel computing for power system climate resiliency: Solving a large-scale stochastic capacity expansion problem with mpi-sppy,”Electric Power Systems Research, vol. 235, p. 110720, Oct. 2024

    work page 2024

  8. [8]

    Climate-Resilient Nodal Power System Expansion Planning for a Realistic California Test Case,

    A. Musselman, T. V . Zuluaga, E. Glista, M. Monteagudo, J. M. Grappone, and J.-P. Watson, “Climate-Resilient Nodal Power System Expansion Planning for a Realistic California Test Case,” Mar. 2025. [Online]. Available: https://optimization-online.org/?p=29697

    work page 2025

  9. [9]

    A scalable solution framework for stochastic transmission and generation planning problems,

    F. D. Munoz and J.-P. Watson, “A scalable solution framework for stochastic transmission and generation planning problems,”Computa- tional Management Science, vol. 12, no. 4, pp. 491–518, Oct. 2015

    work page 2015

  10. [10]

    Assessing the economic value of co-optimized grid-scale energy storage investments in supporting high renewable portfolio standards,

    R. S. Go, F. D. Munoz, and J.-P. Watson, “Assessing the economic value of co-optimized grid-scale energy storage investments in supporting high renewable portfolio standards,”Applied Energy, vol. 183, pp. 902–913, Dec. 2016

    work page 2016

  11. [11]

    Scenarios and Policy Aggrega- tion in Optimization Under Uncertainty,

    R. T. Rockafellar and R. J.-B. Wets, “Scenarios and Policy Aggrega- tion in Optimization Under Uncertainty,”Mathematics of Operations Research, vol. 16, no. 1, pp. 119–147, Feb. 1991

    work page 1991

  12. [12]

    Progressive hedging for stochastic energy management systems,

    V . Kaisermayer, D. Muschick, M. Horn, and M. G ¨olles, “Progressive hedging for stochastic energy management systems,”Energy Systems, vol. 12, no. 1, pp. 1–29, Feb. 2021

    work page 2021

  13. [13]

    S. P. Boyd and L. Vandenberghe,Convex Optimization. Cambridge University Press, Mar. 2004

    work page 2004

  14. [14]

    A parallel hub-and-spoke system for large-scale scenario-based optimization under uncertainty,

    B. Knueven, D. Mildebrath, C. Muir, J. D. Siirola, J.-P. Watson, and D. L. Woodruff, “A parallel hub-and-spoke system for large-scale scenario-based optimization under uncertainty,”Mathematical Program- ming Computation, vol. 15, no. 4, pp. 591–619, Dec. 2023

    work page 2023

  15. [15]

    Pyomo/mpi-sppy,

    D. L. Woodruff, B. Knueven, D. Mildebrath, C. Muir, J. D. Siirola, and J.-P. Watson, “Pyomo/mpi-sppy,” Sep. 2025. [Online]. Available: https://github.com/Pyomo/mpi-sppy

    work page 2025

  16. [16]

    Grid Structural Characteristics as Validation Criteria for Synthetic Networks,

    A. B. Birchfield, T. Xu, K. M. Gegner, K. S. Shetye, and T. J. Overbye, “Grid Structural Characteristics as Validation Criteria for Synthetic Networks,”IEEE Transactions on Power Systems, vol. 32, no. 4, pp. 3258–3265, Jul. 2017

    work page 2017

  17. [17]

    Optimization of Electricity Systems Under Un- certainty,

    T. Valencia Zuluaga, “Optimization of Electricity Systems Under Un- certainty,” Ph.D. dissertation, UC Berkeley, 2024

    work page 2024

  18. [18]

    Roads to Removal: Options for Carbon Dioxide Removal in the United States,

    J. Pett-Ridge, H. Ammar, and A. Aui, “Roads to Removal: Options for Carbon Dioxide Removal in the United States,” Lawrence Livermore National Laboratory, Tech. Rep. LLNL-TR-852901, 2023. APPENDIXA LIST OF SYMBOLS IN OPTIMIZATION MODEL A. Nomenclature Sets B: Set of all buses (nodes) in the network. Indexed byb(or i, m, nwhere noted). G: Set of generation ...

    work page 2023