Numerical simulations which require High-Performance Computing (HPC) are more and more used in all the fields of industry, yielding performance improvement for numerous products.
This trend is similar in sports, especially when material and technology are involved. Numerical simulations become useful tools as performance aids: from weather forecast to estimate snow quality on the skiing track to use the best possible wax to Formula One racing car for which a lot of optimisation (aerodynamics, structure, combustion...) are carried out thanks to great power computer.
Nautical sports are not left behind, especially in sailing, where the design workflow is similar to the one used in the industry, since some of classes can afford skills, tools and computer power, as for the America's cup.
A sport like rowing do not benefit from such a financial support. In addition to that, from the hydrodynamic point of view, it has some specificities which make them challenging: the athlete who is the propulsive machine has a great interaction with the system and uses blades which generate a violent unsteady flow near the free surface. The unsteadiness of the propulsion associated to the motion of the athlete with respect to his hull leads to large secondary motions which are a singular feature for the flow around hulls in calm water in hydrodynamics. Another fluid-structure interaction (FSI) which influences the response of the global system boat-rower(s)-oar(s) is due to the flexibility of the oar shaft.

Previous research works have been done about the validation of the flow around the hull and around the blade ([1]).
Here, a focus is done on a recent work aiming at developping a high-fidelity modelling of the complete boat-oar(s)-rower(s) system. This development starts with the contribution of a group of engineering students who follow the project-based specialisation named "Paris Scientifiques 2024/ Scientific Challenge 2024" at Centrale Nantes.
To achieve this scientific challenge, a multibody system is developed to accurately model the kinematics of the rower with respect to its environment. This imposed kinematics is driven both by some gesture parameters and in-situ data measurements as the time evolution of the sweep angle of each oar. The dynamics of the global system is then reduced to the dynamics of the hull, which is solved by integrating the major fluid forces acting on both the hull and the blades through CFD. This is done by coupling the Navier-Stokes solver ISIS-CFD developed by the METHRIC team at the LHEEA Lab. with a dedicaded Python program which models the multibody system boat-oar(s)-rower(s).
Such a CFD configuration brings into play advanced numerical techniques such as the combined use of overset grid technique and adaptive grid refinement (AGR).
To reach an efficient and robust algorithm for this partionned approach, the coupling iteration occurs during the non-linear iterations of the fluid solver because it is the most costly part. This implicit internal coupling solves both the dynamics of the hull and the flexibility of the oar shafts. Data transfer between the two codes is done through a TCP socket using the ZMQ library. As other fluid-structure interaction in hydrodynamics, a stabilization procedure based on an artificial added mass method is used to tackle the destabilising added-mass effects.

At term, such model targets to bring objective and unbiased criteria for questions which have up to now only empirical answers.

[1] Yoann Robert. Simulation numérique et modélisation d'écoulements tridimensionnels instationnaires à surface libre. Application au système bateau-avirons-rameur. PhD thesis, Centrale Nantes, 2017.
URL :https ://hal.archives-ouvertes.fr/tel-01719696/le/Robert2017.pdf