The NextSim interview series aims to introduce readers to the people behind our research. Vincent Couailler, a numerical methods and fluid mechanics expert and head of the Numerical Methods for Fluid Dynamics group in the department of Aerodynamics, Aeroelascticity and Acoustics of ONERA, is involved in NextSim´s work in algorithms improvement for efficient solution process. Learn more about his thoughts and motivations regarding his work and the NextSim project.

What is your position at ONERA and what is your research area?

I am head of the research unit NFLU (Numerical Methods for Fluid Dynamics) in the department of Aerodynamics, Aeroelascticity and Acoustics of ONERA.

The scope of our research work goes from the design and fundamental analysis of numerical methods for fluid mechanics to their implementation in prototype or finalized software for the production of innovative CFD tools.

More precisely, our activities concern the development of numerical methods representing complex physical phenomena, including turbulent flows, multifluid, multiphysics coupling especially for aeroacosutics and aerothermal systems.

Our research also concerns the field of stochastic numerical models, as well as meta-models for the calculation and analysis of problems associated to huge data management, with an increasing investment in algorithms using artificial intelligence tools.

In our unit we have strong links with the French and international academic world, especially in the context of EU projects and thesis supervision. Within ONERA we are responsible for the new generation CFD software CODA, developed in cooperation with DLR and AIRBUS.

The activities that we carry out in NextSim with the project partners to improve the efficiency and the robustness of classical and high-accurate methods on unstructured and hybrid meshes are perfectly consistent with our research themes.

You are involved in the work package on algorithms improvement for efficient solution process. Could you explain why this is important for NextSim?

ONERA is leading the wokpackage “Algorithms improvement for efficient solution process”, a topic which is of main importance for the success of CODA. Robust and efficient integration methods of space-time accurate scheme are the mathematical and numerical basis to get very efficient HPC codes. It is also crucial to ensure that their specification and their implementation or re-engineering are carried out jointly with massively parallel architecture specialists who are present in the NextSim project.

What would you have done if you were not a researcher?

My career as a researcher in aerospace results from my education in mathematics at Sorbonne University, the desire to work in a field involving a strong activity in applied mathematics, and the exciting challenges of aeronautical and space discovery which fascinated me during my youth ( Concorde, Apollo). I could have worked in another application area, such as econometrics. I have also always been very interested in history, where a professional career could have interested me as well. I am very happy to have chosen this professional domain, CFD, combining mathematics, physics and computer science, with ambitious and stimulating challenges with industrial partners.

Have you encountered any challenges in pursuing your research career?

In a professional career, and particularly in the field of research, a certain number of activities having to reconcile very different scientific and organizational objectives can be considered as challenges. I will take as an example a technical and strategic challenge that I had in the 90s: Develop in less than 2 months a new version of a CFD code ensuring the stability and reliability of the results on complex applications, impossible to achieve with the previous version, for a major aircraft manufacturer. This was linked to a GO/NOGO decision for the continuation of an emerging cooperation in the field of major codes of fluid mechanics. The development consisted of choosing a new discretization approach, rewriting the data structure and associated numerical schemes of a multiblock 3D code, and evaluating it on a complex industrial aircraft configuration. Fortunately, it was a success. Even if not all the challenges are successful, they always lead to going further if they are defined with coherent and measurable objectives.

How is your experience with NextSim so far?

NextSim is a very stimulating project both for the scientific and technical objectives and for the dynamic management and cross-interactions thanks to a clear planning of several technical and review meetings. The project is a special opportunity to integrate new developers with different profiles for the HPC improvement of CODA, also to extend the interest for these next-generation solvers within the European community. Due to the pandemic, it has not been possible to organize a physical meeting but I hope it will be possible soon because such interactions strengthen the cooperation between the teams.

What do you hope NextSim can achieve beyond the life of the project?

I hope NextSim will establish a nucleus of a European community for the development of the next generation CODA software, a lasting kernel that will grow over the next few years. I also hope that this project will establish a stronger link between specialists in numerical analysis and specialists in computer science.

What advice would you give young researchers who would like to follow in your footsteps?

I would say that as a researcher, you are motivated by investigating the specific field you have chosen by contributing to either new knowledge or new finalized tools useful to the whole community, or by both. In a context of very dynamic but very evolving digitalization, it is necessary to maintain and develop expertise beyond today’s more complex environments.