Efficient control with Microsoft Fabric optimizes hydrogen production at OGE

Green hydrogen is seen as a key to an efficient energy transition. Surplus wind power is used in its production. This requires transparent digital mapping of the process chains. So far, this has been a challenge.

OGE therefore uses the Microsoft Fabric digitalization platform to present hydrogen production as a digital twin. Real-time production data is analyzed in order to make data-based decisions for operations.

The platform links real-time electricity prices with internal data and calculates the expected price for 1 kg of H₂. This allows the production plan to be optimized. This reduces manufacturing costs by up to 50%.

The future of power supplies depends on sustainable sources of energy. Gas grid operator Open Grid Europe (OGE) is already doing pioneering work at its own Krummhörn compressor station and demonstrating how surplus wind power can be used to produce green hydrogen. Efficient control of that production calls for transparent processes. That’s why OGE relies on a data platform based on Microsoft Fabric. And the platform even goes one step further by using real-time data to optimize production efficiency.

This solution, which is based on Microsoft Fabric, enables us not only to evaluate consolidated data, but also to make data-driven decisions for operations and how to control them.

Ralf Werner, CIO, Open Grid Europe GmbH

The challenge: Isolated data makes connecting processes difficult

The German government's goal is for Germany to be climate-neutral by 2045. In order to achieve this, sustainable, renewable energy sources such as wind, solar, or green hydrogen must be developed and increasingly used for electricity, mobility, and industry. However, relying on renewable energy sources for the power supply poses challenges: “Green power generation will always be volatile, including in the future,” explains Ralf Werner, CIO at Open Grid Europe (OGE). “If there’s no wind, no sun, and we can’t generate electricity from renewables, then we need some kind of backup supply. This is where we see the greatest potential for green hydrogen.”

That is what led the gas network operator OGE to launch the KRUH2 project at the Krummhörn compressor station, on Germany’s North Sea coast, where it operates an electrolyzer using surplus renewable electricity generated by onshore and offshore wind turbines. The electrolyzer produces green hydrogen, which OGE in turn uses as an energy source for its own site.

The major obstacle OGE faced was that the process chains for hydrogen production couldn’t be mapped digitally end to end. “But that’s an absolutely essential step,” Werner adds, “because we have to know how much H₂ is needed and will be purchased. Only then can we produce the right quantities and adapt our production planning to the requirements and options for interim storage.”

That calls for data—and for the right technology to organize the data and ultimately analyze it. With this in mind, OGE is working with a start-up called CibusCell, which it first came into contact with via the hydrogen innovation hub H2UB. Their joint goal is to use Microsoft Fabric to digitally map the processes of hydrogen production from end to end within the special requirements of a system-critical environment.

The solution: Digital twin enables optimum hydrogen production

OGE now uses a digitalization platform to map the entire hydrogen production process: the platform collects real-time data from the hardware components, evaluates it, and maps it as a digital twin. “This solution, which is based on Microsoft Fabric, enables us not only to evaluate consolidated data, but also to make data-driven decisions for operations and how to control them,” Werner says.

But the solution can do even more than that: it pairs data on internal production costs with real-time data on electricity prices from the European Energy Exchange. This integration in turn allows the price of producing one kilogram of H₂ to be calculated in real time, so that the platform can use it to make suggestions for optimizing the production plan. That means electricity can be purchased cost-effectively, thereby reducing production costs.

More about this diagram

To fully exploit the data’s potential, Microsoft Azure and Machine Learning were also integrated into the platform. And this has had a significant impact, as Marcus Rübsam, Managing Director of CibusCell Technology, explains: “We can bring down production costs by around 50% just through economic optimization based on machine learning.”

We can bring down production costs by around 50 percent just through economic optimization based on machine learning.

Marcus Rübsam, Managing Director, CibusCell Technology GmbH

Werner adds: “Microsoft isn’t only a technology provider for us, but has also supported us in handling our software requirements. Through a private offer via Azure Marketplace, we were able to commission CibusCell directly.”

Since OGE, as a regulated company, is not allowed to use the hydrogen it produces commercially, it is used exclusively at the Krummhörn compressor station, at hydrogen filling stations for the company’s in-house vehicle fleet and for the site’s heat and energy supply. "With KRUH2, we have nevertheless done real pioneering work. The project is not only a proof of concept for green hydrogen as an energy source of the future, but also for how sector coupling can be successful," explains Ralf Werner. "In addition, the development of KRUH2 has enabled us to gather valuable knowledge to provide answers to future questions. This will enable us as a company to play an even more active role in the development of a Germany-wide hydrogen transport network and help drive forward the energy transition in Germany."