This study proposes a mixed-integer second-order cone optimization model for the cost-optimal siting, sizing, and operation of electrolyzers in integrated hydrogen and medium-voltage electricity networks, including cost-optimal network expansion. Unlike studies based on standardized IEEE test cases, the model is applied to a real-world representative Dutch distribution system using detailed DSO data on grid topology and electricity and hydrogen demand and supply forecasts for 2024 and 2035. The case study focuses on the increasingly congested Oost-Groningen region in The Netherlands Optimal electrolyzer deployment is assessed across 45 scenarios varying hydrogen price, hydrogen demand, and electrolyzer CAPEX. While deployment patterns differ by scenario and year, several no-regret sites consistently emerge near large solar parks. Results show that distributed electrolyzers can eliminate grid congestion and associated reinforcements in 2024, and reduce reinforcement costs by up to 60% in 2035, when reinforcements remain necessary to integrate renewables and additional electrolyzer load. Coupling the local hydrogen grid to the national network is critical for electrolyzer profitability and effective congestion relief. Overall, the results support deploying electrolyzers in congested distribution grids, as this substantially reduces grid reinforcement costs. Results provide actionable insights for European DSOs facing electricity and hydrogen infrastructure planning. • A MISOCP model jointly optimizes electrolyzer siting, sizing, operation, and energy grid expansion. • Model applied to real electricity and hydrogen grids using Dutch DSO data for 2024 & 2035. • Distributed electrolyzers fully resolve 2024 congestion and cut 2035 reinforcement costs by up to 60%. • Sensitivity analysis identifies no-regret electrolyzer locations near major solar generation sites. • Connecting the regional hydrogen grid to the national backbone important for electrolyzer profitability and congestion relief.
Koenen et al. (Fri,) studied this question.