Simulações de correntes de turbidez bi-dispersas na configuração lock-release com valores diferentes de Schmidt

Abstract

Density current flows is the phenomenon of one fluid propagate through another fluid with a different density due to the gradient of density. Turbidity currents is the name for flows when this difference is caused by suspended sediments due to the turbulent flow. The study of this kind of flow own great relevance for academic and economic issues. For example, there is a great interest for the industry of oil and gas to understand better this phenomenon, because deposits create by sedimentation of turbidity currents flows with a great concentration of fine particles have a great potential to become hydrocarbon reservoirs. There are many kinds of approaches for the study of density current flows, one of them is through the computational simulation by a numerical scheme. This work is a numerical study of bi-disperse density currents flows on lock-release configuration. The goal is understand the effect of Reynolds and Schmidt number variation on the flow dynamic as well as the deposit formation. Simulations using direct numerical simulation approach as well as implicit large eddy simulations are performed. The code Incompact3D was used to solve the Navier-Stokes equations and transport equation under the Boussinesq approximation. Values between 5000 and 70000 are performed for the Reynolds number. For the Schmidt number, two cases are considered: (i) 1 for the two particles fractions and (ii) value 3 for the particle fraction with bigger grain diameter, while we keep unitary value for the other particle fraction. In order to quantify our experiments we calculated some features of the flow as position and velocity of front, suspended mass, sedimentation rate and deposit height. We also calculated the temporal evolution of the energy budget of our simulations. Our results agree with literature. Analyzed the effect of Reynolds number, it notices that the simulations with greater values of Reynolds are more closely to reference. For the double mass difusiv- ity effect, we see it is more significant for simulations with Reynolds 5000 and decreasing in relevance for bigger values of Reynolds.