OPmobility automates Hydrogen storage system design process with AI

As the mobility sector shifts towards sustainable propulsion systems, hydrogen emerges as a promising energy vector, particularly for heavy-duty and long-range vehicles. However, the integration of hydrogen storage systems poses a critical design challenge: how to safely and efficiently embed multiple high-pressure tanks within highly constrained vehicle architectures. Traditional design approaches, often manual and iterative, struggle to meet the performance, cost, and packaging constraints required by the market.

‍

To address this complexity, OPmobility partnered with Dessia Technologies to develop an AI solution capable of automating pressure vessel layout design. This collaboration brought together object-oriented system modeling, rule-based engineering, and advanced AI algorithms to transform the way hydrogen storage architectures are designed.

‍

Hydrogen storage: Context and constraints

‍

Hydrogen storage via compressed gas, typically at 700 bar, requires the use of Type IV composite tanks with polymer liners. These tanks offer weight and safety advantages but are geometrically constrained: they must maintain specific shapes to limit material stress under high pressure. As a result, their integration into a vehicle becomes a multidimensional optimization problem, where available space, thermal performance, mechanical constraints, and cost-efficiency all intersect.

‍

‍

OPmobility, a global leader in hydrogen mobility, has strategically positioned itself across the hydrogen value chain, from tanks and fuel cells to complete storage systems. For OPmobility, the ability to design and validate layout configurations rapidly and systematically is key to improving RFQ responsiveness, optimizing cost-performance trade-offs, and scaling hydrogen system production.

‍

‍

Inside the design process: Why change was needed?

‍

For design engineers, translating these requirements into an actual hydrogen vessel architecture wasn’t straightforward. The process was slow, manual, and often frustrating. It relied heavily on CAD tools, design intuition, and repeated trial-and-error to explore even a few viable configurations. With limited time to respond to RFQs or test alternatives, design teams were often forced to settle for what was “good enough.”

‍

The rigid shape of high-pressure tanks made the challenge even harder. Fitting them inside a constrained vehicle volume, without compromising safety or efficiency, was like solving a puzzle with fixed pieces and no room for flexibility. Every iteration required time-consuming modeling and revalidation. And because decisions weren’t well-documented or repeatable, it was difficult to explain why one configuration was chosen over another or to reuse knowledge in future projects.

‍

What OPmobility needed wasn’t just faster tools, it was a smarter way of working. A system that could automate design generation, handle complexity, and give engineers the ability to explore a wide range of architectures with full confidence in every result.

‍

The solution: Automated architecture generation with AI

‍

Using Dessia’s AI libraries, OPmobility deployed a custom application based on the following key pillars:

‍

1. Object-oriented system modeling

‍

Each hydrogen storage system is represented as a modular object, with attributes like size, shape, mounting constraints, and spatial limits. The overall system is defined as a hierarchy of components and relationships, enabling flexible yet rigorous modeling of the vehicle volume and tank positioning logic.

‍

‍

2. Explainable AI & rule libraries

‍

The layout generation engine is driven by a symbolic AI framework. Engineering rules — such as minimum distances, symmetry preferences, stacking logic, center-of-gravity constraints — are encoded directly into the system. This ensures every design variant complies with structural, spatial, and safety requirements.

‍

‍

3. Generative algorithms

‍

Rather than iterating manually, the application explores the full solution space:

• All combinations of tanks from a predefined catalogue
• All valid arrangements within the 3D integration envelope
• All configurations meeting defined technical constraints

‍

This generative process runs in minutes and can output hundreds of valid layout scenarios, each documented and ready for comparison.

‍

‍

‍

4. Statistical AI for decision support

‍

Once solutions are generated, clustering and dimensionality reduction techniques are applied to organize designs into architecture families. This enables the engineering team to:

• Compare design trade-offs across cost, volume, and feasibility
• Focus only on high-potential configurations
• Justify selections with quantitative evidence

‍

Beyond architecture generation, Dessia’s AI solution can also connect with existing CAD and PLM systems. This allows design engineers to import geometry, export selected configurations, and manage design data within the tools they already use.

‍

‍

What changed for OPmobility

‍

The implementation of AI-driven architecture generation brought immediate and measurable improvements to OPmobility’s hydrogen storage design process.

‍

Design time was significantly reduced, what previously took several days of manual CAD work could now be completed in just a few hours. Instead of exploring only a few configurations, design engineers could access a wide range of valid architecture options, all generated within the constraints of the system.

‍

This broader exploration was backed by rule-based validation, which helped ensure that every configuration was technically feasible from the start. As a result, OPmobility gained greater confidence in the selected designs and was able to respond more quickly to RFQs with well-documented, technically sound proposals.

‍

Conclusion

‍

This project shows how AI-driven design tools can bring a new level of speed, precision, and robustness to engineering. By automating the generation of high-pressure hydrogen tank architectures, OPmobility turned a slow, manual process into a fast and intelligent design engine.

‍

But it’s not just about saving time, it’s a real strategic advantage in an industry racing to integrate hydrogen systems efficiently and competitively. Dessia’s AI libraries have proven to be a reliable and scalable foundation for intelligent design exploration, built to meet industrial needs and ready to support the future of hydrogen mobility.

‍