A Cournot-Nash Model for a Coupled Hydrogen and Electricity Market

Pavel Dvurechensky, Caroline Geiersbach, Michael Hintermüller, Aswin Kannan, Stefan Kater, Gregor ZöttlarXiv 2024 · DOI · arXiv

Abstract: We present a novel model of a coupled hydrogen and electricity market on the intraday time scale, where hydrogen gas is used as a storage device for the electric grid. Electricity is produced by renewable energy sources or by extracting hydrogen from a pipeline that is shared by non-cooperative agents. The resulting model is a generalized Nash equilibrium problem. Under certain mild assumptions, we prove that an equilibrium exists. Perspectives for future work are presented.

Introduction: The Challenge of Variable Renewable Energy

As global energy systems transition toward variable renewable energy (VRE) sources like wind and solar, the primary challenge becomes balancing supply and demand. Unlike traditional fossil fuel plants, VRE output is intermittent. "Power-to-Gas" (PtG) technology—converting excess electricity into hydrogen for storage in existing gas pipelines—is a leading solution for long-duration energy storage.

However, modeling the market dynamics of such a coupled system is complex. Electricity markets operate on near-instantaneous timescales, while gas transport in pipelines is a relatively slow physical process governed by Partial Differential Equations (PDEs). This research addresses the gap by creating a unified game-theoretic framework where non-cooperative firms compete across both networks.

Conceptual illustration of a coupled hydrogen and electricity infrastructure featuring renewable generation, transmission lines, and gas conversion stations.

Problem Formulation and Market Physics

The core of the paper is the formulation of a Generalized Nash Equilibrium Problem (GNEP). In this model, firms strategically decide their levels of electricity generation, the volume of electricity to convert into hydrogen for storage, and the amount of energy to sell.

Unlike a standard Nash equilibrium, a GNEP accounts for the fact that players' action sets are interdependent due to shared physical constraints. For instance, the total amount of hydrogen injected into a pipeline by all firms cannot exceed the physical pressure limits of that pipe. If one firm increases its injection, it physically restricts the capacity available to others.

Methodology: Bridging Time Scales and Physics

To make the coupling mathematically tractable, the authors employ several key modeling techniques:

  • PDE-Constrained Optimization: The hydrogen network is modeled using a linearized version of the ISO2 model for isothermal gas flow. This allows the model to track pressure and mass flow across space and time.
  • DC Load Flow Approximation: A standard linear approximation is used for the electricity grid to maintain computational feasibility while respecting transmission limits.
  • Multiscale Integration: The model operates on an intraday scale, treating electricity as continuous-time to align with the continuous nature of gas transport PDEs.
  • Graph Abstraction: The physical infrastructure is mapped onto a graph where nodes represent generation points, sale points, and conversion stations (PtG and GtP).
Graph representation of the coupled network model showing the interaction between electricity nodes and hydrogen nodes.

Key Insights: The Pipeline as a Battery

A significant insight of this work is the formalization of "line packing." By increasing the pressure within the gas pipeline, the infrastructure itself acts as a dynamic storage device—essentially a massive, distributed battery. This allows firms to "store" energy during periods of high renewable production and low prices, then extract it later via Gas-to-Power (GtP) stations when demand peaks.

To ensure the existence of a stable market equilibrium, the authors linearized the gas flow equations. While this simplifies the underlying physics, it provides the mathematical concavity required for game-theoretic stability. Without this, the model could become a non-convex problem where a stable market outcome might not exist.

Theoretical Results and Existence Proof

The primary contribution of the paper is Theorem 3.6, which provides a formal proof for the existence of a generalized Nash equilibrium in this coupled system. Using the Nikaido-Isoda function and fixed-point theory, the authors demonstrate that under "mild" assumptions—such as convex cost functions and bounded capacities—there exists a set of strategies where no firm can unilaterally increase its profit.

Furthermore, the authors prove that the underlying system of gas flow PDEs is uniquely solvable and depends continuously on the boundary conditions (the decisions made by the firms), ensuring the market model is well-posed.

Limitations and Future Research

While the mathematical foundation is robust, the authors identify several areas for future development:

  1. Nonlinear Physics: The current model relies on linearized gas equations. Moving to fully nonlinear, non-convex models would better reflect real-world physics but presents significant challenges for proving equilibrium existence.
  2. Information Asymmetry: The model assumes "perfect information," where all agents know the costs and strategies of their competitors. Future work will explore distributed versions of the game with limited information sharing.
  3. Numerical Implementation: This paper is theoretical; the authors plan to develop specific algorithms (such as extragradient-type methods) to calculate these equilibria in practical, large-scale scenarios.
  4. Dynamic Topology: The current framework assumes a fixed network. Future iterations could incorporate long-term investment decisions, such as where to optimally place new PtG stations to maximize grid stability.

Summary updated

Abstract

We present a novel model of a coupled hydrogen and electricity market on the intraday time scale, where hydrogen gas is used as a storage device for the electric grid. Electricity is produced by renewable energy sources or by extracting hydrogen from a pipeline that is shared by non-cooperative agents. The resulting model is a generalized Nash equilibrium problem. Under certain mild assumptions, we prove that an equilibrium exists. Perspectives for future work are presented.
BibTeX
@misc{dvurechensky2024cournot,
  author = {Pavel Dvurechensky and Caroline Geiersbach and Michael Hintermüller and Aswin Kannan and Stefan Kater and Gregor Zöttl},
  title = {{A Cournot-Nash Model for a Coupled Hydrogen and Electricity Market}},
  year = {2024},
  eprint = {2410.20534},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2410.20534},
}
Figure