We have interviewed David Le Touzé, Deputy Director of LHEEA Lab (ECN and CNRS), as part of our strong partnership with Ecole Centrale de Nantes.
Nextflow Software is pushing the limits of simulation with next-generation CFD software, thanks to its team and also based on more than 10 years of close partnership with leading academic research laboratories, especially with Ecole Centrale de Nantes.

By watching David Le Touzé video testimony (3 minutes), you will discover our genesis and the SPH method evolution.

Hello, my name is David Le Touzé.
I’m professor of fluid mechanics at Ecole Centrale Nantes, French Grande école of Engineering. I’m mainly expert on Computational Fluid Dynamics (CFD) and especially on the Smoothed Particle Hydrodynamics (SPH) method.

Your best memory or anecdocte?

When I just finished my PhD, I was heading towards a post doc in Italy and before I left, my former PhD director told me “This method will never work”.

 

Why do you work on the SPH method?

So we started to work on this method about 20 years ago. At the time, it was very difficult to model with a conventional CFD complex air-water interface. It was also very difficult to have in fluid dynamics simulation multiple bodies or multiple objects which we were moving together.
The SPH method was quite attractive because it actually permitted to do these things quite straightforwardly, but of course it was a very young method, so we had to do lots of academic developments.

Once it was mature, of course, the next step has been to industrialize the method, so it was the role of Nextflow Software. Nextflow Software today is applying the method to a number of industries, many industrial problems. Especially in the automotive industry, like flows in gearbox, hydroplaning flows, cooling of electric motors. But also outside the automotive industry, helicopter or aircraft ditching and emergency landing on water, big waves on the ship, industrial processes, and even in health, engineering for health now.

 

Your CFD and meshless methods vision?

So there are two aspects for me in this question.
The first one is the meshing. The meshing is the operation that takes a lot of time to engineers. It’s a more lengthy operation in terms of time, and this is of course linked to the mesh, which you don’t have in meshless methods. What is for sure is that, in the future, this meshing operation, it will be transparent. Whatever you have behind, this will be transparent to the user.
The second aspect is the ratio between cost and efficiency. If we want to achieve a really good ratio between accuracy and cost, we need, for me, to have many couplings or fusions or merging of methods.

 

Your last word

So it’s only a start for SPH.

___

David Le Touzé

Deputy Director, LHEEA Lab (ECN and CNRS)