Hydrogen fuel cells: Clean energy for the transportation sector

May 12, 2025 fuel cell technology 10 min read

The development of hydrogen fuel cells is highly complex and cannot be effectively managed with conventional methods and hardware-based prototype testing. However, Model-Based Design can help overcome these challenges.

Electric vehicles powered by hydrogen fuel cells (HFCs) as a clean, CO2-free means of propulsion are not a new concept. Although battery-powered electric vehicles dominate the roads, partly due to the simpler charging infrastructure setup, hydrogen requires not only significant energy to produce but also a network of hydrogen fueling stations, which is currently lacking.

Nonetheless, HFC systems have potential, albeit less so in personal transportation. Their advantages include a lower vehicle weight compared to batteries, higher power density, quicker refueling, and longer range. They are also suitable for continuous operation, making HFC drives highly appealing for heavy-duty trucks, buses, trains, maritime ships, and even airplanes (Figure 1). Application tests are currently underway in all these areas, facilitated by the fact that hydrogen fueling stations can be more easily centralized in these sectors, eliminating the need for a widespread infrastructure.

Figure 1: The use of hydrogen fuel cells will increase significantly by 2050, especially in seagoing vessels. © Segula Technologies

Challenges in HFC Development

Polymer Electrolyte Membrane Fuel Cells (PEM-FCs), also known as Proton Exchange Membrane Fuel Cells, are most commonly used in the applications mentioned above. However, modeling them poses a significant challenge for engineers. Despite the seemingly simple concept, PEM-FCs are highly complex in their overall dynamics. They consist of a cell stack with electrodes and a membrane, as well as components like tanks, pressure reducers, compressors, and elements for moisture and thermal management. All these components must be dimensioned, optimized, coordinated, and tested to ensure maximum efficiency and performance over a long lifespan.

The dynamics of the actual cell stack, which involve electrochemical processes and transport phenomena of gases, water vapor, and water droplets, are particularly complex. Conventional development methods and hardware-based prototype testing are insufficient to handle the sheer number of parameters that require optimization.

Segula Technologies relies on Model-Based Design to develop customized HFC applications for the mentioned industries. The company uses MATLAB®, Simulink®, Simscape™, and AI methods as the foundation. Simscape is used to model physical systems within the Simulink environment.

The Simscape Model as a Starting Point

One of the starting points for HFC development at Segula is the Simscape fuel cell model from MathWorks (Figure 2). This includes all the peripheral components mentioned above, as well as the cell stack. It features a custom domain created in Simscape that dynamically captures the dynamics of all four involved gases: nitrogen, oxygen, hydrogen, and water vapor. This is crucial for optimizing both the performance and longevity of the cell stack. With the help of this model alone, the engineers were able to shorten the initial development phase by four to six weeks.

Figure 2. The complete system designed by Segula in the automotive version. The Simscape model of the actual fuel cell is highlighted in the blue box.
© Segula Technologies.

Segula engineers chose to use and further develop this model because they were particularly interested in the transport phenomena within the cell stack. The MathWorks model already includes the dynamics of water vapor transport and nitrogen accumulation on the anode side, both of which were further refined.

Additionally, the flow of water droplets within the system and the associated heat flow were considered to further increase the model’s accuracy. The physics of the cell interior was represented with 3D modeling tools, allowing for precise capture of water and heat flow. The data was then integrated into the original Simscape model. With these captured dynamics, subsequent simulations at the system level were accelerated and made more flexible.

Application Range of the Dynamic Model

The extended, more accurate simulation model of the physical HFC system can be used in various ways. Initially, it is used for component selection and dimensioning. Subsequently, parameters and control strategies can be optimized. Control strategies can also be designed and validated against the model as a plant model, such as for energy generation, moisture and thermal management, pressure control, or the regular expulsion of accumulated nitrogen.

Finally, energy flows between the buffer battery, fuel cell, and drive can be analyzed, and the vehicle’s range can be determined based on different driving profiles. All this occurs without building physical prototypes, eliminating waiting times and associated costs. Additionally, a broader range of operating conditions can be tested, even in extreme scenarios, without risking hardware damage or overload.

A significant advantage of the created model is its flexibility in any hydrogen fuel cell application, whether for road, rail, sea, or air transport. The model accurately determines if component dimensions are correct, cells deliver the required performance, and control functions as desired. “The interaction of individual HFC components, tested and optimized through modeling and simulation, is crucial for the energy efficiency and longevity of the fuel cell,” emphasizes Dirk Rensink, Technical Lead for Fuel Cell Simulation at Segula.

AI-Based Expert System for Variant Parameterization

Given that Segula engineers obtained limited data from real hardware, they developed an AI-based expert system to evaluate previous simulation runs. This AI extracts parameters from the recorded data that can be applied to variants and entirely new configurations of the HFC model. With this approach, the team can avoid starting from scratch for each new fuel cell system and estimating the best values for new configurations. This not only saves time in the initial phase but also allows the models themselves to evolve as the amount of available data grows.

Early Validation and Hardware-in-the-Loop Simulation

Using the models, the Segula team can test the design of a controller before assembling a prototype.

Traditionally, a system prototype is developed on a test bench, and the control software is then tested and calibrated (Figure 3).

Figure 3: A fuel cell stack on the test bench. © Segula Technologies

In the model-based approach, the Simscape model of the fuel cell system is used to generate code that is loaded into a real-time simulation computer. This is known as Hardware-in-the-Loop (HIL) simulation, and the model serves as an environment for testing the fuel cell software used in real systems. Fuel cell controllers are tested and validated with HIL tests to simulate the behavior of the fuel cell during a typical week of operation or even a 30,000-hour service life. The main focus of the tests is on the controllers for the compressor and the humidification system.

“With these pre-tests, we can calibrate the correct values, and we were very close to the real values. Starting the system on the test bench with the calibrated, pre simulated model instead of from scratch significantly accelerates the development time,” explains Stephan Schnorpfeil, Head of the Fuel Cell Team at Segula.

Together with the AI-supported parameter database, the system models lead to a shorter time-to-market for customers. Based on the developed models and acquired expertise, Segula can also offer its customers specialized solutions that would hardly be feasible without simulations. The system’s flexibility enables knowledge transfer for the development of PEM-FCs for a wide range of applications and performance requirements. For example, a ship fuel cell can be adapted for an automotive application. Customers have also used the resources built by Segula to optimally dimension components for their own PEM-FCs.

Synergies of Theory and Practice: More Efficient Hydrogen Synthesis

Some of the fuel cells developed in this way are already in use. As a result, the Segula team increasingly has access to field data, enabling them to further refine their models. Simulation and real operation are increasingly converging through this ever-expanding database, improving the foundation for data- and model-based development of hydrogen fuel cell systems.

The physics behind fuel cells and hydrogen electrolyzers is comparable in essential aspects, as both technologies are based on similar electrochemical principles. Applying the presented results to electrolyzers for hydrogen production and adapting the Simscape model from MathWorks or the refined Segula model could potentially lead to efficiency improvements in the field of renewable energies within the mobility sector. The goal of producing and using hydrogen in a carbon-neutral and economically feasible manner may thus be one step closer.

About the authors

Vijayalayan R heads automotive industry and location field application engineering at MathWorks India. He and his team facilitate the adoption of Model-Based Design, revolutionizing engineering processes and steering the industry towards innovative solutions. His efforts are focused on empowering customers to embrace these next-generation technologies in their journey towards electrification, AI, and softwaredefined vehicle projects. As the secretary of the SAE India Bengaluru section and member of their management committee, Vijayalayan’s influence extends beyond MathWorks, impacting the broader automotive industry.