Modeling Electrolyzers and Gas Networks for Integration with Power Systems

The massive integration of electricity generation from intermittent renewable energy sources tremendously increases the need for flexibility in the power system. Additionally, the decarbonization of the gas system requires the replacement of natural gas. Hydrogen produced from renewable electricity by electrolyzers that are part of hybrid power plants or energy islands is envisioned as a solution to both problems. In recent years, natural gas-fired power plants have mainly provided the flexibility needed to balance increasing intermittent renewable power production. Their operational space to provide flexibility can be enlarged by supplying the required amount of natural gas through pipeline networks. This requires coordination in the operation of power and gas systems. Missing coordination may increase system operational costs and require costly network management measures, e.g., load shedding, in extreme situations. The energy system operation should further consider the nonlinear physics of components and energy flows in networks. Improving coordination and representation of nonlinear physics is critical to harnessing the flexibility of multi-energy systems.

This thesis analyzes how electrolyzers and gas networks can be appropriately modeled for integration with power systems to unlock their full flexibility potential. Three modeling levels are identified based on their scope and detail: component, system, and policy.

On the component level, two novel relaxation-based models that accurately capture the nonlinear relationship between power consumption and hydrogen production of the electrolyzer are proposed. The first model is a linear relaxation of the state-of-the-art piecewise linearization approach. In contrast, the second model is a conic relaxation of a quadratic approximation. It is proven mathematically that the conic relaxation is exact under prevalent operating conditions. The models are numerically evaluated for the optimal scheduling of a hybrid power plant, considering a realistic case study. Both models outperform the state-of-the-art approach. The conic model extends the operational space of the electrolyzer compared to the state-of-the-art model representation. It is shown that adequately capturing the nonlinear physics of the electrolyzer fosters its cross-carrier flexibility, increasing hydrogen production and economic profitability.

On the system level, a unified framework for analyzing modeling and solution choices of the partial differential equations governing gas flow in pipelines is proposed. Existing approaches in the literature are derived within the framework, enabling a rigorous comparison. The assumption of ex-ante known gas flow directions is challenged based on the uncertain demand of gas-fired power plants that balance fluctuations from intermittent renewable power production. Using a realistic case study, it is analyzed how incorrect predetermination of gas flow directions in the optimization model reduces the gas network flexibility by limiting the operational space of gas-fired power plants. Based on two stylized and realistic large-scale case studies, the impact of different modeling and solution choices of the gas flow physics on the gas network flexibility is evaluated. Accurately modeling of linepack is needed to exploit the intertemporal flexibility from slow gas flow transients. Furthermore, it is shown how relaxation-based models overestimate the available linepack flexibility. Modeling the gas flow directions and physics in the operation of integrated power and gas systems is essential to unlock the flexibility potential that gas networks can provide to power systems.

On the policy level, an idealized co-optimization of day-ahead energy and real-time balancing markets is considered. It is modeled using a two-stage stochastic optimization problem representing the uncertainty from intermittent renewable power production via scenarios in the balancing market stage. The flexibility of electrolyzers on energy islands to compensate for fluctuations in renewable power supply in the balancing market is analyzed. Furthermore, it is investigated how the bidding zone allocation of the energy island impacts this flexibility. Based on the developed model, electrolyzers on energy islands are found to have low capacity factors and provide little balancing flexibility in the market. Two policy recommendations are derived to enhance the flexibility potential. First, energy islands should constitute their own bidding zone to more transparently reflect structural congestion in the electricity grid. Second, financial support may incentivize flexibility provision in the balancing market.