Researchers at Delft University of Technology present a mathematical planning tool to help hydrogen-focused companies run their distributed assets and take part in several energy markets at once. The paper proposes a centralized, portfolio-level optimization that jointly decides how to operate electrolyzers, renewable generators, and storage while choosing when and where to buy and sell electricity, hydrogen, and green certificates.

The team built the model as a mixed-integer linear program (MILP), a standard optimization method that can handle on/off decisions and linear constraints. The model sees all company assets at once and co-optimizes day-ahead scheduling and market bids for wind and solar farms, electrical energy storage (EES), electrolyzers for producing hydrogen, and downstream hydrogen uses. It supports participation in the electricity market, physical and virtual Power Purchase Agreements (PPAs, assumed here as “take-as-produced”), bundled and unbundled hydrogen sales, and trading of Guarantees of Origin (GOs), the green certificates tied to renewable output.

At a high level, the approach unlocks flexibility by coordinating assets across sites instead of treating them one by one. For example, electricity from on-site renewables, PPAs, or the day-ahead market can be routed to electrolysis, to storage, or sold, depending on prices and company policy targets. The model also enforces company-level rules such as green hydrogen targets and additionality requirements, so market choices and technical schedules meet internal compliance needs.

The authors tested the model on three operational scenarios to see how different compliance strategies change outcomes. In those cases, centralized portfolio control increased hydrogen production by 2.42 times and lowered daily operational costs by 9.4% while meeting the paper’s company-level policy constraints. The study also examines how different PPA dispatch formats and policy constraints affect costs and production.