No abstract

SUMMARYWe present a finite element residual‐based variational multiscale formulation applied to the numerical simulation of particle‐laden flows. We employ a Eulerian–Eulerian framework to describe the flows in which the mathematical model results from the incompressible Navier–Stokes equation combined with an advection–diffusion transport equation. Special boundary conditions at the bottom are introduced to take into account sediments deposition. Computational experiments are organized in two examples. The first example deals with the well‐known gravity current benchmark, the lock‐exchange configuration. The second also employs for the current initiation the lock configuration, but the sediment particles are endowed with a deposition velocity and are allowed to leave the domain in the moment they reach the bottom. This is intended to mimic, partially, as the bed morphology is not allowed to change, the deposition process, in which sediment deposits are no longer carried by the flow. The spatial pattern of the deposition and its correlation with flow structures are the main focus of this analysis. Numerical experiments have shown that the present formulation captures most of the relevant turbulent flow features with reasonable accuracy, when compared with highly resolved numerical simulations and experimental data. Copyright © 2013 John Wiley & Sons, Ltd.

No abstract

The exploitation of oil and gas can be stimulated through hydraulic fractures (HF), which are discontinuities in the rock formation induced by the injection of high pressurized viscous fluids. Because there exists considerable variability in geologic formations, such as oil and gas reservoirs, the computational models, and, consequently, the predictions drawn from simulations, might lead to misleading conclusions, despite the use of efficient and robust numerical schemes. In order to take into account uncertainties on the numerical results due to the variability in the input data, a stochastic analysis of HF is presented here. The elasticity modulus of the rock and the confining stress are assumed to be described by random variables, and therefore, the equations governing the fracture propagation are recast as stochastic partial differential equations (SPDEs). In order to solve the resulting problem, among several alternatives available in the literature, a stochastic collocation method is adopted. The elasticity modulus probability distributions are constructed using two different approaches, both using a small amount of information. A number of numerical simulations are presented in order to illustrate the impact of the uncertainties in the data input on the fracture propagation.