This work presents a state-of-the-art computational approach for performing large-scale SIMULATIONs of turbulence on heterogeneous compute nodes. The approach proposed combines message passing interface (MPI) parallelization for distributing the computation across nodes with open accelerators (OpenACC) to target a graphical processing unit (GPU) acceleration. Two different OpenACC approaches, corresponding to managed and non-managed data management, have been tested and compared. The results obtained indicate that the non-managed strategy outperforms the non-managed scheme in all cases. Additionally, the non-managed GPU-accelerated approach presents an overall speedup of 6× with respect to the CPU version. Finally, the capability of the resulting parallel and accelerated flow solver to perform high-fidelity SIMULATIONs of turbulent flow is validated against a canonical wall-bounded fluid problem.

Simulating fluid flow at the micro level is an ongoing problem. Simplified macroscopic flow models like Darcy’ s law is unable to estimate the fluid dynamic properties of porous media. The digital sample reconstruction by high-resolution X-ray computed tomography scanning and fluid-dynamics SIMULATION, together with the increasing power of super-computers, allow to carry out pore-scale SIMULATIONs through digitally-reconstructed porous samples. The pore-scale flows which derived from computational fluid dynamics are then evaluated using the finite volume method implemented in the open-source platform OpenFOAM® . In this work, to verify the solver in porous media, we simulated fluid flow around a sphere in body-centered cubic (bcc) lattice and calculated the dimensionless permeability for a wide range of radius and porosity; the results are comparable with those obtained by using Carman-kozeny equation. Then this solver is performed on a realistic sample to investigate the effect of sample size on calculated permeability and tortuosity and the mesh refinement levels for a fixed image resolution.

An overview of removal and re-entrainment of particles in turbulent flows is presented. The procedure for the DIRECT NUMERICAL SIMULATION (DNS) of the Navier-Stokes equation via a pseudospectral method for simulating the instantaneous fluid velocity field is described. Particle removal mechanisms in turbulent flows in a duct are examined and effects of the near-wall coherent eddy on the particle detachment and removal are discussed. The particle equation of motion, including the hydrodynamic drag and lift, as well as the Brownian and gravitational forces, are used in the analysis. The SIMULATION results show that large size particles move away roughly perpendicular to the wall due to the action of the shear lift force. Small particles, however, follow the upward flows formed by the near wall eddies in the low speed streak regions. Thus, turbulent near wall eddies play an important role in small particle resuspension, while the lift force is an important factor for re-entrainment of large particles. The results also suggest that small particles primarily move away from the wall in the low speed streaks, while larger particles are mostly removed in the high-speed streaks.        

The objective of this paper presents the results of subtask dealing with the bond behaviour study of the reinforcement systems under monotonic loading pull-out tests. This NUMERICAL method is based on the slip and the bond stress distributions through the anchored length of the bar in the concrete block. The work refers, especially to the implementation of reinforcing bars and bond-slip models between steel and concrete in the developed finite element program. For the application of the proposed method, three analytical expressions of bond-slip relationship are selected. The obtained results are presented and commented with the fundamental characteristics of plain concrete and reinforced concrete members. The bond models in contribution with concrete and reinforcing steel provide a relativity good representation of bond-zone system responses.