For hydrogen to become the flexible clean fuel for heavy industry and shipping, hydrogen developers need to reduce production costs while having the capacity and ability to scale dynamically as required. Stable, cost-effective hydrogen production is as central to developing the hydrogen economy as creating the demand and logistics network needed to support it.

Developments in automation, particularly machine-learning-enabled AI models, have led to an industrial revolution over the past five years. Reimagining the legacy machine design and commissioning process, as well as traditional onboarding and training, and standard operating procedures, with automation and machine learning as a core tenet, can create a new standard of cost and efficiency.

Hydrogen developers can future-proof their operations by making informed planning decisions to integrate simulation and machine learning

One opportunity for hydrogen production developers is to scale from concept to commercial production faster. To make costs competitive, producers need commercial deployments to be of a specific size that meets their needs. As demand increases, producers also need to expand capacity. This can be challenging due to supply issues with the requisite equipment; currently, lead times for some hardware can stretch up to two years. To support flexibility, open-architecture, standards-based automation hardware design can allow producers to integrate components and systems from multiple suppliers, reducing sourcing risk.

Green hydrogen and blue hydrogen production systems have unique needs and often application-specific designs that can present a substantial systems integration challenge. The design phase of commercial projects is a considerable portion of a plant’s lifetime costs. Modern design for automation leverages simulation and machine learning to trial the installation, commissioning and operation of new equipment designs entirely digitally, identifying hardware and software challenges before construction is completed. This reduces unexpected downtime and substantially accelerates the commissioning process.

Simulation streamlines everything

Simulation for machine design has become a familiar tool in many industrial settings. Beyond identifying issues and facilitating commissioning, design simulation drives a systems-based approach to create adaptable, modular designs that enable dynamic scaling. Machine-learning–powered simulation has expanded its role beyond planning to enable high-efficiency operations and facilitate effective onboarding programmes that can help organisations develop the skillsets for hydrogen production that are currently lacking in the labour market.

The simulations that enable developers to test machines virtually also allow employees to train in real-world situations using granular simulations. Unique simulation scenarios can be devised to manage risk and increase safety through the simulation and planning of varying weather and process conditions. A high-fidelity simulation of real-world scenarios means planning procedures and training for high-risk processes can be integrated into operations from the start, rather than being developed after the fact.

Once operations are underway, simulation has become an essential tool for operational efficiency, from monitoring specific parts for predictive and prescriptive maintenance to optimising entire supply chains. AI-empowered resource planning can adapt production to demand and unexpected changes in the supply chain. This can include maximising the benefits of onsite renewable energy generation to maintain the lowest possible cost per kilogram produced. Cross-departmental functions, such as assignment scheduling, logistics and maintenance, can be optimised to maximise plant uptime. Simulation using real-time data can also optimise CO₂ capture processes for blue hydrogen plants.

The basis of robust simulation is real-time data that encompasses the breadth of the hardware and software sources. This unified data architecture enables real-time simulation for process optimisation and retrospective high-level analysis, allowing for continuous cost competitiveness. The architecture is facilitated by a software-abstracted standards-based design that itself requires a similar unification and standardisation of control software. This design foundation, where information from every sensor, machine and planning system is standardised and unified in a data repository, also forms the core of meeting future sustainability certification standards for hydrogen production. The holistic lifecycle accounting of electricity sources, CO₂ capture rates, leakage controls and storage guarantees will require real-time, unified data and detailed process analytics to provide a trace for every kilogram of hydrogen produced, to meet those guidelines.

Commercial hydrogen horizon

The hydrogen economy is evolving rapidly. There are numerous demand-side pilot projects ranging from on-site industrial power generation to shore-based power for ships at port. Hundreds of orders for LNG-fuelled vessels are being placed annually, and vessels powered by hydrogen and hydrogen-derived ammonia are entering service. Hydrogen and ammonia bunkering solutions for ports are being tested or are underway—including those at Amsterdam, Rotterdam, Klaipeda and IJmuiden.

As the first of the demand-side projects are expected to come online in 2026, such as the cruise ship Viking Libra, demand will continue to scale up through 2030. After 2030, multiple commercial corridors for hydrogen are expected to begin stabilising demand. The growing demand ecosystem for hydrogen at an industrial scale will facilitate larger commercial-scale hydrogen production. Commercial-scale production should create a demand floor for equipment vendors and lower the cost of parts for producers as suppliers increase their production capacity.

After 2030, multiple commercial corridors for hydrogen are expected to begin stabilising demand

To remain competitive in this emerging market, hydrogen producers need to be able to optimise production to minimise their costs. Simulation-driven design reduces overengineering and identifies issues before they create downtime and unexpected expenses. Predictive and prescriptive maintenance extends the life of components. The success of hydrogen production will rely as much on software and data integration for production optimisation and tracking as on physical infrastructure and demand. These toolsets can help market leaders bridge the transition where demand will be uncertain and operations need to extract the most value possible from every component and production hour. Simulation and machine learning provide the framework to support producers as the stable foundation of demand is built.

The 2025 changes to the 45V tax credits for hydrogen producers require projects to begin construction by 2028 to remain eligible under the revised law. This has accelerated deployment timelines for developers. The long lead times for industrial equipment, sometimes up to two years, mean developers who want to take advantage of 45V have to be making production decisions now. With hydrogen demand already scaling, developers must ensure operations can meet the rising demand on schedule. The emerging need for on-demand power generation in datacentres is creating a new and rapidly growing market for hydrogen solutions.

As leaders in an emerging market, hydrogen production requires close controls on cost, scope and operations. Conversely, as demand emerges, these same producers must be able to ramp up production and scale operations to meet the new demand requirements. Hydrogen developers can future-proof their operations by making informed planning decisions to integrate simulation and machine learning, designing for automation from the outset, and utilising standards-based, open-architecture hardware that facilitates interoperability. These decisions help minimise costs and facilitate agile operations.

David Meyer is Sustainable Energy Solutions Leader, Siemens Industry

---

Author: David Meyer