Suchuan Dong

A combined direct numerical simulation-particle image velocimetry study of the turbulent near wake

We investigate the near wake of a cylinder at values of Reynolds number corresponding to the onset and development of shear-layer instabilities. By combining quantitative experimental imaging (particle image velocimetry, PIV) and direct numerical simulations at Re = 3900/4000 and 10 000, we show that the flow structure is notably altered. At higher Reynolds number, the lengths of both the wake bubble and the separating shear layer decrease substantially. Corresponding patterns of velocity fluctuations and Reynolds stress contract towards the base of the cylinder. The elevated values of Reynolds stress at upstream locations in the separated layer indicate earlier onset of shear-layer transition. These features are intimately associated with the details of the shear-layer instability, which leads to small-scale vortices. The simulated signatures of the shear-layer vortices are characterized by a broadband peak at Re = 3900 and a broadband high spectral-density ‘plateau’ at Re = 10 000 in the power spectra. The shear-layer frequencies from the present direct numerical simulations study agree well with previous experimentally measured values, and follow the power law suggested by other workers.

Direct numerical simulation of turbulent Taylor-Couette flow

We investigate the dynamical and statistical features of turbulent Taylor–Couette flow (for a radius ratio 0.5) through three-dimensional direct numerical simulations (DNS) at Reynolds numbers ranging from 1000 to 8000. We show that in three-dimensional space the Gortler vortices are randomly distributed in banded regions on the wall, concentrating at the outflow boundaries of Taylor vortex cells, which spread over the entirecylinder surface with increasing Reynolds number. Gortler vortices cause streaky structures that form herringbone-like patterns near the wall. For the Reynolds numbers studied here, the average axial spacing of the streaks is approximately 100 viscous wall units, and the average tilting angle ranges from 16° to 20°. Simulationresults have been compared to the experimental data in the literature, and the flow dynamics and statistics are discussed in detail.

A time-stepping scheme involving constant coefficient matrices for phase-field simulations of two-phase incompressible flows with large density ratios

We present an efficient time-stepping scheme for simulations of the coupled Navier-Stokes Cahn-Hilliard equations for the phase field approach. The scheme has several attractive characteristics: (i) it is suitable for large density ratios, and numerical experiments with density ratios up to 1000 have been presented; (ii) it involves only constant (time-independent) coefficient matrices for all flow variables, which can be pre-computed during pre-processing, so it effectively overcomes the performance bottleneck induced by variable coefficient matrices associated with the variable density and variable viscosity; (iii) it completely de-couples the computations of the velocity, pressure, and the phase field function. Strategy for spectral-element type spatial discretizations to overcome the difficulty associated with the large spatial order of the Cahn-Hilliard equation is also discussed. Ample numerical simulations demonstrate that the current algorithm, together with the Navier-Stokes Cahn-Hilliard phase field approach, is an efficient and effective method for studying two-phase flows involving large density ratios, moving contact lines, and interfacial topology changes.

Flow past a trapezoidal tab

The flow passing over a trapezoidal tab mounted on a flat plate is studied using direct numerical simulation (DNS). Such a tab has been used to generate hairpin-like vortices to enhance cross-stream mixing. We attempt to provide a detailed account of the three-dimensional topology and dynamics of the hairpin vortices in the tab wake. Simulations are conducted for three tab inclination angles at Reynolds number Re=600 based on the free-stream velocity and the tab height. A finite-volume discretization scheme involving 2.6 × 10 6 control volumes is employed for the simulations and the results are compared with PIV experimental data. Simulations captured all the experimentally observed near-field flow features including a pair of streamwise corotating vortices and its transition to hairpin vortices. Simulation results provide new insight into the vortex dynamics in the tab wake

NEKTAR, SPICE and Vortonics: using federated grids for large scale scientific applications

In response to a joint call from USs NSF and UKs EPSRC for applications that aim to utilize the combined computational resources of the US and UK, three computational science groups from UCL, Tufts and Brown Universities teamed up with a middleware team from NIU/Argonne to meet the challenge. Although the groups had three distinct codes and aims, the projects had the underlying common feature that they were comprised of large-scale distributed applications which required high-end networking and advanced middleware in order to be effectively deployed. For example, cross-site runs were found to be a very effective strategy to overcome the limitations of a single resource. The seamless federation of a grid-of-grids remains difficult. Even if interoperability at the middleware and software stack levels were to exist, it would not guarantee that the federated grids can be utilized for large scale distributed applications. There are important additional requirements for example, compatible and consistent usage policy, automated advanced reservations and most important of all co-scheduling. This paper outlines the scientific motivation and describes why distributed resources are critical for all three projects. It documents the challenges encountered in using a grid-of-grids and some of the solutions devised in response.

A robust and accurate outflow boundary condition for incompressible flow simulations on severely-truncated unbounded domains

We present a robust and accurate outflow boundary condition and an associated numerical algorithm for incompressible flow simulations on unbounded physical domains, aiming at maximizing the domain truncation without adversely affecting the flow physics. The proposed outflow boundary condition allows for the influx of kinetic energy into the domain through the outflow boundaries, and prevents un-controlled growth in the energy of the domain in such situations. The numerical algorithm for the outflow boundary condition is developed on top of a rotational velocity-correction type strategy to de-couple the pressure and velocity computations, and a special construction in the algorithmic formulation prevents the numerical locking at the outflow boundaries for time-dependent problems. Extensive numerical tests for flow problems with bounded and semi-bounded physical domains demonstrate that this outflow boundary condition and the numerical algorithm produce stable and accurate simulations on even severely truncated computational domains, where strong vortices may be present at or exit the outflow boundaries. The method developed herein has the potential to significantly expedite simulations of incompressible flows involving outflow or open boundaries, and to enable such simulations at Reynolds numbers significantly higher than the state of the art.

Turbulent flow between counter-rotating concentric cylinders : a direct numerical simulation study

We report three-dimensional direct numerical simulations of the turbulent flow between counter-rotating concentric cylinders with a radius ratio 0.5. The inner-and outer-cylinder Reynolds numbers have the same magnitude, which ranges from 500 to 4000 in the simulations. We show that with the increase of Reynolds number, the prevailing structures in the flow are azimuthal vortices with scales much smaller than the cylinder gap. At high Reynolds numbers, while the instantaneous small-scale vortices permeate the entire domain, the large-scale Taylor vortex motions manifested by the time-averaged field do not penetrate a layer of fluid near the outer cylinder. Comparisons between the standard Taylor-Couette system (rotating inner cylinder, fixed outer cylinder) and the counter-rotating system demonstrate the profound effects of the Coriolis force on the mean flow and other statistical quantities. The dynamical and statistical features of the flow have been investigated in detail.

Cross-site computations on the TeraGrid

The TeraGrids collective computing resources can help researchers perform very-large-scale simulations in computational fluid dynamics (CFD) applications, but doing so requires tightly coupled communications among different sites. The authors examine a scaled-down turbulent flow problem, investigating the feasibility and scalability of cross-site simulation paradigms, targeting grand challenges such as blood flow in the entire human arterial tree.

Dual-level parallelism for high-order CFD methods☆

A hybrid two-level parallel paradigm with MPI/OpenMP is presented in the context of high-order methods and implemented in the spectral/hp element framework to take advantage of the hierarchical structures arising from deterministic and stochastic CFD problems. We take a coarse grain approach to OpenMP shared-memory parallelization and employ a workload-splitting scheme that reduces the OpenMP synchronizations to the minimum. The hybrid algorithm shows good scalability with respect to both the problem size and the number of processors for a fixed problem size. For the same number of processors, the hybrid model with 2 OpenMP threads per MPI process is observed to perform better than pure MPI and pure OpenMP on the SGI Origin 2000 and the Intel IA64 Cluster, while the pure MPI model performs the best on the IBM SP3 and on the Compaq Alpha Cluster. A key new result is that the use of threads facilitates effectively p-refinement, which is crucial to adaptive discretization using high-order methods. � 2003 Elsevier B.V. All rights reserved.

BDF-like methods for nonlinear dynamic analysis

We present several time integration algorithms of second-order accuracy that are numerically simple and effective for nonlinear elastodynamic problems. These algorithms are based on a general four-step scheme that has a resemblance to the backward differentiation formulas. We also present an extension to the composite strategy of the Bathe method. Appropriate values for the algorithmic parameters are determined based on considerations of stability and dissipativity, and less dissipative members of each algorithm have been identified. We demonstrate the convergence characteristics of the proposed algorithms with a nonlinear dynamic problem having analytic solutions, and test these algorithms with several three-dimensional nonlinear elastodynamic problems involving large deformations and rotations, employing St. Venant-Kirchhoff and compressible Neo-Hookean hyperelastic material models. These tests show that stable computations are obtained with the proposed algorithms in nonlinear situations where the trapezoidal rule encounters a well-known instability.