3 questions with Florent Brissaud – NaTran

Florent Brissaud joined NaTran in 2018, where he currently serves as R&D activity leader and expert reference for functional safety. Prior to that, he worked as a consultant for a large company and then independently. He has been using GRIF since his consulting years, particularly the Tree, Petri, and Petro modules, and appreciates the scientific rigor, modeling power, and user-friendliness of these tools. Today, GRIF allows him to model feared events on gas transmission network facilities and test various maintenance policies to improve reliability while controlling costs.

NaTran is the main operator of high-pressure gas transmission in France. Formerly known as GRTgaz, it changed its name in 2025 to affirm its commitment to the energy transition. NaTran now transports gases that are central to the energy challenges of tomorrow: natural gas, biomethane, hydrogen, and CO₂.

We also use Tree extensively for managing our industrial assets. We leverage data from our CMMS to estimate equipment reliability, accounting for aging (Weibull laws), and we develop “virtual age” models to assess the impact of maintenance on aging.

These models were then integrated into Tree and the Boolean module, in co-development, to generalize our approach across various facilities. We use this especially to model feared events on our 4,500 gas delivery stations. These stations handle gas pressure reduction, metering, overpressure prevention, and supply assurance. We must avoid supply interruptions and overpressure for availability and safety reasons.

We have developed tools to assess equipment reliability based on age and characteristics. Then we build fault trees for each feared event (supply interruption, overpressure) depending on the station’s architecture (single or dual line, number of regulators, valves, safety devices, etc.). With 2 feared events and 4,500 stations, we have 9,000 fault trees to create. So, we developed a tool to generate these fault trees based on station architecture, equipment, and characteristics. Code lines were written to call Tree to compute all the generated fault trees in bulk, then compile all results into a CSV file. This system lets us identify the highest-risk stations, test different maintenance policies (current or alternative), and optimize at the network level.

Until recently, we applied a uniform maintenance policy. Thanks to this approach, we can justify differentiated maintenance, adapting preventive maintenance intervals based on the specific risks of each facility.

What makes GRIF a unique tool in your opinion?

It’s a tool we have used for a long time and know well. As an R&D project manager, what matters to me is scientific rigor—I trust the quality of the calculations, without questionable approximations. That’s the main argument for me.

Then there’s the modeling power—we have everything we need, especially with the Boolean and Simulation packages. We often started with simulations using Petri (e.g., virtual age models), which allowed us to validate our theoretical models.

Once validated, these models were integrated into Tree to improve computation times and benefit from a tool we were already using. We also appreciated the collaboration with the user club and Satodev, which enabled the integration of features we needed. As an R&D center, we often go beyond what the tool offers, so we value being able to integrate our own models.

Finally, an important point: the ergonomics. I find the tool very well designed, especially for checking or modifying parameters globally, and for simulating or validating models quite easily.