Multiphase computational fluid dynamics for subsea jet trenching applications

Details

Supervisors

Legat, Vincent ; Lambrechts, Jonathan ; Constant, Matthieu

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil mécanicien, à finalité spécialisée

Abstract

Subsea jet trenching is used to create trenches and bury cable in the sea floor, to bring electricity from offshore wind farms to the continent, for example. This is a preliminary study to determine if jet trenching can be modelled numerically and give accurate results. This is done using the software Migflow that solves immersed granular flows, and the results are compared to the laboratory experiments done by A.T.H. Perng and H. Capart \cite{perng_underwater_2008}, where a small vehicle uses jets to create a trench in a tank. The Navier-Stokes equations and the incompressibility hypothesis are already implemented in the software. To those equations is added the concentration equation, that gives the evolution of the concentration of water with time. That equation is the sum of an advection equation and a mass diffusion equation. The sedimentation velocity, velocity at which the sand settles on the sea floor, is also determined, tested and added to the concentration equation. This software, modified with the new equations, is tested on the setup done by Perng and Capart. The tests are first done to determine the impact on the simulations of the viscosity of the sand, the coefficient of sedimentation, the mass diffusion coefficient and the height of the jet. Finally, the velocity of the vehicle is analysed and compared to the laboratory results. The flux at the jet is also tested. The conclusions are that the current model do not recreate the results of the experiments. Indeed, the results obtained when changing the velocity of the vehicle do not match the ones obtained in the laboratory. This is due to the dependence of the solution to the time step and to the use of the wrong model for the viscosity of the sand. This last problem could be solved modelling the sand as a non-Newtonian fluid. Despite the results to be improved, the concentration equations and the sedimentation velocity are both validated and implemented correctly in the software.