Jeffrey Wright

Second-order upwind and central difference schemes for recirculating flow computation

Two-dimensional driven cavity flows with the Reynolds number ranging from 102 to 3.2 x 10 3 are used to assess the performance of second-order upwind and central difference schemes for the convection terms. Three different implementations of the second-order upwind scheme are designed and tested in the context of the SIMPLE algorithm, with the grid size varying from 21 x 21 to 161 x 161 uniformly spaced nodes. Converged solutions are obtained for all Reynolds numbers. Although these different implementations of the second-order upwind scheme have the same formal order of accuracy, significant differences in numerical accuracy are observed. It is demonstrated that better performance can be obtained for the second-order upwind scheme if the discretization is cast in accordance with the finite volume formulation. Although both the second-order upwind and central difference schemes exhibit no oscillations in the solution, the upwind scheme is more accurate. In assessing and comparing the performance of these schemes, the distribution of cell Reynolds number is discussed and its impact on numerical accuracy illustrated.

Development of pressure-based composite multigrid methods for complex fluid flows

Abstract Progress in the development of a multiblock, multigrid algorithm, and a critical assessment of the κ-e two-equation turbulent model for solving fluid flows in complex geometries is presented. The basic methodology employed is a unified pressure-based method for both incompressible and compressible flows, along with a TVD-based controlled variation scheme (CVS), which uses a second-order flux estimation bounded by flux limiters.Performance of the CVS is assessed in terms of its accuracy and convergence properties for laminar and turbulent recirculating flows as well as compressible flows containing shocks. Several other conventional schemes are also employed, including the first-order upwind, central difference, hybrid, second-order upwind and QUICK schemes. For better control over grid quality and to obtain accurate solutions for complex flow domains, a multiblock procedure is desirable and often a must.Here, a a composite grid algorithm utilizing patched (abutting) grids is discussed and a conservative flux treatment for interfaces between blocks is presented.A full approximation storage-full multigrid (FAS-FMG) algorithm that is incorporated in the flow solver for increasing the efficiency of the computation is also described. For turbulent flows, implementation of the κ-e two-equation model and in particular the wall functions at solid boundaries is also detailed.In addition, different modifications to the basic k-e model, which take the non-equilibrium between the production and dissipation of κ and e and rotational effects into account, have also been assessed.Selected test cases are used to demonstrate the robustness of the solver in terms of the convection schemes, the multiblock interface treatment, the multigrid speedup and the turbulence models.

Validation of a New Parallel All-Speed CFD Code in a Rule-Based Framework for Multidisciplinary Applications

This paper focuses on the validation of a new all-speed Computational Fluid Dynamics (CFD) code called LociSTREAM. This computational package is not just another CFD solver; rather, it integrates proven numerical methods and state-of-the-art physical models to compute all-speed flows using generalized grids in a novel rule-based programming framework called Loci which allows: (a) seamless integration of multidisciplinary physics in a unified manner, and (b) automatic handling of massively parallel computing. The objective is to be able to routinely simulate problems involving complex geometries requiring large unstructured grids with arbitrary polyhedral cells and complex multidisciplinary physics. As a first step towards achieving this objective, a wide range of model test cases are studied here, including incompressible laminar and turbulent flow cases, inviscid compressible flow cases, compressible turbulent flows with wall heat transfer as well as internal turbulent flows in 3D geometries and unsteady computations. Comparison of the code with experimental and prior benchmark numerical results is done to validate the robustness of the code for flows ranging from incompressible to supersonic regimes. A scalability analysis is performed as well to study the efficiency of parallelization.

Turbulence-Chemistry Interaction and Heat Transfer Modeling of H 2 /O 2 Gaseous Injector Flows

Reliable prediction of rocket injector flows introduces significant challenges associated with the complex physics involving recirculation, turbulence, scalar mixing, chemical reactions and wall heat transfer. This work is aimed at assessing the importance of turbulence-chemistry interaction and non-equilibrium effects in experimentally characterized single and multi-element injector flows. By examining the different chemistry models (laminar finite rate, assumed PDF with either flamelet or equilibrium assumption), it was found that for both cases investigated, chemical non- equilibrium is insignificant while substantial turbulence-chemistry interaction is observed. A zonal wall treatment was developed based on a blend of SST low-Re turbulence wall treatment and law- of-the-wall, showing improved predictive capability. A heat flux extraction method was also proposed to estimate heat flux results from adiabatic flamelet model under the consideration that wall heat loss is small compared to the overall energy generated by chemical reactions.

Convective film cooling over a representative turbine blade leading-edge

Abstract Computations are performed to simulate a discrete hole film cooling flow over an experimental test geometry representative of the leading edge of turbine blades. A multiblock pressure correction algorithm is used for the computations, and both low-Reynolds number and wall function k – e models are used for turbulence closure. The flow through the coolant ducts, from the plenum to the blade surface, is resolved as a part of the computation by specifying the coolant mass flux in the plenum. A systematic grid refinement study is conducted with the finest grid consisting of approximately one million points. Next, the flowfield is examined ; key physical mechanisms resulting from the interactions between the cooling jets and the freestream are identified and their effect on the thermal field is compared with the experimentally observed thermal field. Finally, a study of geometric parametric variation is conducted to optimize the film cooling design. Nine different combinations of two parameters, namely, the relative stagger and the relative angle between the two rows of cooling holes are investigated for their effect on heat transfer on the blade surface.

Pressure-based viscous flow computation using multiblock overlapped curvilinear grids

A pressure-based multiblock computational method is developed for solving the incompressible Navier-Stokes equations in general curvilinear grid systems. A conservative interface scheme is devised with desirable accuracy to handle the information transfer between blocks. The scheme is based on the semiimplicit-type flow solver with the staggered grid. Issues concerning discontinuous grids, global mass conservation, viscous term treatment, and boundary conditions at the grid interface are addressed. The method is tested for two flow problems, a curved channel flow and a bifurcated channel flow. The calculations demonstrate that, besides maintaining desirable solution characteristics across discontinuous grid interfaces, the present multiblock algorithm can achieve convergence rates comparable to that of the single-block algorithm, yielding an improved computational capability for treating complex flow problems.

An Algorithm for Chemically Reacting Flows on Generalized Grids Using a Rule-Based Framework

This paper reports on the development of a new CFD tool called Loci-STREAM. This tool is not just another CFD solver; rather, it integrates proven numerical methods for generalized grids and state-of-the-art physical models in a novel programming framework called Loci which allows: (a) seamless integration of multidisciplinary physics in a unified manner, and (b) automatic handling of massively parallel computing. The objective is to be able to routinely simulate problems involving complex geometries requiring large unstructured grids and complex multidisciplinary physics. An immediate application of interest is simulation of reacting flows in combustion devices, especially rocket engines.

An All-Speed Solver for Unsteady Reacting and Non-Reacting Flows Using a Rule-Based Framework

A CFD solver called Loci-STREAM is presented in this paper. This solver integrates proven numerical methods for generalized grids and state-of-the-art physical models in a rule-based programming framework called Loci which allows: (a) seamless integration of multidisciplinary physics in a unified manner, and (b) automatic handling of massively parallel computing. The objective is to be able to routinely simulate problems involving complex geometries requiring large unstructured grids and complex multidisciplinary physics. Practical applications include flows in combustion devices, turbomachinery, etc.

Computation of a leading-edge film cooling flow over an experimental geometry

Computations are performed to simulate a leading edge film cooling flow over an experimental test geometry. An in-house CFD–heat transfer code based on a pressure correction algorithm is used for the computations. The code allows the use of multiple blocks in the domain with discontinuous grid lines while maintaining flux conservation at block interfaces. A second–order TVD–based controlled variation scheme (CVS) is used for discretization, along with k–e models with options of using wall functions or low–Reynolds number modifications. From the viewpoint of incorporating CFD into the design process with fast turn–around times, the approach taken in this study is to attempt to simulate the key features of the flowfield with a reasonable grid size, preferably consisting of no more than 250,000 grid points. In order to attain the desired accuracy with these constraints, effective combinations of grid distribution, discretization operators and turbulence models are investigated, and the sensitivity of the computed solution to these factors examined. The results agree qualitatively with the experimental data though some notable quantitative differences can be observed. An attempt is made to explain the key features of the flowfield resulting from the interaction of coolant jets with the hot freestream.Copyright

Numerical simulation of unsteady convective intrusions in a thermohaline stratification

The time-dependent development of convective intrusions in a thermohaline stratification is investigated using a composite grid method with local refinement. An imposed constant sidewall heat flux causes convective cells to form at the heated wall, which subsequently propagate into the bulk domain. For the composite grid computational method, a grid interface treatment based on strict local flux conservation alone allows arbitrary, nonphysical jumps in the temperature and salinity across block boundaries for certain boundary conditions. A revised treatment based upon a linear interpolation with conservative correction is employed to overcome this difficulty. Detailed features of the internal intrusion structure are captured and the sharp interfaces between neighboring intrusions are handled with fine resolution. Two classes of thermohaline intrusions corresponding to recently-obtained experimental results have been computed. Comparison with these results shows excellent agreement both qualitatively, in terms of the internal structure of the intrusions, and quantitatively, in terms of the physical size of the intrusions and the intrusion front propagation speeds. Fundamental differences in the merging processes for the two intrusion classes which were observed in the experiments have also been observed in the numerical results.