Qinling Li

Large-eddy simulation of twin impinging jets in cross-flow

The flow-field beneath a jet-borne vertical landing aircraft is highly complex and unsteady. large-eddy simulation is a suitable tool to predict both the mean flow and unsteady fluctuations. This work aims to evaluate the suitability of LES by applying it to two multiple jet impingement problems: the first is a simple twin impinging jet in cross-flow, while the second includes a circular intake. The numerical method uses a compressible solver on a mixed element unstructured mesh. The smoothing terms in the spatial flux are kept small by the use of a monitor function sensitive to vorticity and divergence. The WALE subgrid scale model is utilised. The simpler jet impingement case shows good agreement with experiment for mean velocity and normal stresses. Analysis of time histories in the jet shear layer and near impingement gives a dominant frequency at a Strouhal number of 01, somewhat lower than normally observed in free jets. The jet impingement case with an intake also gives good agreement with experimental velocity measurements, although the expansion of the grid ahead of the jets does reduce the accuracy in this region. Turbulent eddies are observed entering the intake with significant swirl. This is in qualitative agreement with experimental visualisation. The results show that LES could be a suitable tool when applied to multiple jet impingement with realistic aircraft geometry.

DNS of an oblique shock wave impinging upon a turbulent boundary layer

A supersonic boundary layer at M ∞ = 2, with inflow displacement-thickness Reynolds number Reδ* = 3775, is subjected to an impinging oblique shock. The spatially developing boundary layer is generated using an idealized unsteady analytic inflow profile that emulates the dynamical features of wall-bounded turbulence; this approach has the advantage of creating a self-contained simulation with deterministic inflow conditions that prompt the realistic transfer of energy from the mean flow to the turbulence, and thereby a realistic fully developed turbulent boundary layer in a fairly short downstream distance. The impinging shock induces a small separation bubble and significant intrinsic compressibility effects unrelated to mean property variations.

LES of Impinging Jet Flows Relevant to Vertical Landing Aircraft

The flow-field due to multiple impinging jets from a vertical landing aircraft is highly complex and unsteady. The ability to predict both the mean flow and unsteady excursions is important for the design and development of future aircraft. A Large Eddy Simulation (LES) technique is used to compute two simplified twin impinging jet in cross flow problems. The LES solutions give good agreement with experiment and are a significant improvement over Reynolds Averaged Navier-Stokes solutions. The interaction of the wall jets to form an unsteady fountain that can sometimes reach the intake is observed. There is a disparity in time scales between that needed to resolve the smallest eddies and that for flow to travel through the domain, resulting in the computation needing a large number of time steps to achieve statistically meaningful results. The LES results indicate considerable promise for this type of flow problem, and work is underway to extend the application to full aircraft geometries.

Numerical investigations of wall-bounded turbulence:

This article investigates numerically the effects on turbulence in two important flow regimes – fully developed channel flow and flow past a NACA 0012 airfoil, using the commercial software – FLUENT 6.3. The solution accuracy is explored via a sensitivity study of mesh type and quality effects, employing different element types (e.g. quadrilateral and triangular). The significance of this article is to elucidate the effects of enhancement wall treatment and standard wall function on the turbulent boundary layer. Furthermore, three different turbulence models have been utilized in this study (k−ε, re-normalization group (RNG), and shear stress transport (SST) k−ω). The numerical solutions have been compared with available direct numerical simulation (DNS) and experimental data and very good correlation has been achieved. In addition, the statistical turbulence results related to the RNG turbulence model are shown to yield much closer correlation with DNS and experimental data. The effect of Reynolds number (Re τ = 590 and Re τ = 2320) is studied for the channel flow regime. The near wall resolution is examined in detail by controlling in the y + value. A particularly important objective in this study is to highlight the importance of validation in computational fluid dynamics (CFD) turbulence simulations and sustaining a high degree of accuracy, aspects which are often grossly neglected with industrial CFD software. The authors therefore hope to provide some guidance to applied aerodynamicists utilizing CFD in future studies.

Large eddy simulations for fan-OGV broadband noise prediction

The work in this paper forms part of a project on the use of large eddy simulation (LES) for broadband rotor-stator interaction noise prediction. Here we focus on LES of the flow field near a fan blade trailing edge. The first part of the paper aims to evaluate LES suitability for predicting the near-field velocity field for a blunt NACA-0012 airfoil at moderate Reynolds numbers (2× 105 and 4× 105). Preliminary computations of turbulent mean and root-mean-square velocities, as well as energy spectra at the trailing edge, are compared with those from a recent experiment.1 The second part of the paper describes preliminary progress on an LES calculation of the fan wakes on a fan rig.2 The CFD code uses a mixed element unstructured mesh with a median dual control volume. A wall-adapting local eddy-viscosity sub-grid scale model is employed. A very small amount of numerical dissipation is added in the numerical scheme to keep the compressible solver stable. Further results for the fan turbulentmean and RMS velocity, and especially the aeroacoustics field will be presented at a later stage. Copyright

Grid-Refined LES Predictions for Fan-OGV Broadband Noise

This work forms part of a project on the use of large eddy simulation (LES) for broadband rotor-stator interaction noise prediction. In this paper, we focus on LES calculations of noise sources on and close to a blade trailing edge. We consider two test cases; one an isolated NACA0012 airfoil in flow, and the other an industry-standard rotating fan. In the first case, turbulent mean and RMS velocities and energy spectra at different locations are compared with those from experiment.1,2The sound generated by the unsteady pressure fluctuations on the airfoil surface and by the flow turbulence will be predicted using a Ffowcs Williams Hawkings (FW-H) surface. In the second case, unsteady flow and acoustic fields around the blade passage3 are presented for a refined mesh, and the rotor-stator tonal noise will be predicted by using the rotor-wake mean velocity profile and the methodology described in Lloyd & Peake4. Copyright

Numerical simulation of blowing ratio effects on film cooling on a gas turbine blade

This article investigates the film cooling effectiveness and heat transfer in three regimes for a film-cooled gas turbine blade at the leading edge of the blade with 45 0 angle of injection. A Rolls Royce blade has been used in this study as a solid body with the blade cross section from Hub to Shroud varying with a degree of skewness. A 3-D finite-volume method has been employed (FLUENT 6.3) with a   k turbulence model. The numerical results show the effectiveness cooling and heat transfer behavior with increasing injection blowing ratio BR (1, 1.5 and 2). In terms of the film cooling performance, high BR enhances effectiveness cooling on pressure side and extends the protected area along the spanwise direction from hub to shroud. The influence of increased blade film cooling can be assessed via the values of Nusselt number in terms of reduced heat transfer to the blade.

Fundamental Issues, Technology Development, and Challenges of Boiling Heat Transfer, Critical Heat Flux, and Two-Phase Flow Phenomena with Nanofluids

ABSTRACT This paper presents a comprehensive and critical review of studies on nucleate pool boiling heat transfer, flow boiling heat transfer, critical heat flux (CHF), and two-phase flow phenomena with nanofluids. First, general analysis of the available studies on the relevant topics is presented. Then, studies of physical properties of nanofluids are discussed. Next, boiling heat transfer, CHF phenomena, and the relevant physical mechanisms are explored. Finally, future research needs have been identified according to the review and analysis. As the first priority, the physical properties of nanofluids have a significant effect on the boiling and CHF characteristics but the lack of the accurate knowledge of the physical properties has greatly limited the studies. Fundamentals of boiling heat transfer and CHF phenomena with nanofluids have not yet been well understood. Flow regimes are important in understanding the boiling and CHF phenomena and should be focused on. Two-phase pressure drops of nanofluids should also be studied. Furthermore, economic evaluation of the enhancement technology with nanofluid should be considered for the new heat transfer enhancement technology with nanofluids. Finally, applied research should be targeted to achieve an enabling practical heat transfer and CHF enhancement technology for engineering application with nanofluids.

On the Eect of Convective Flux Formulation for LES of Compressible Flows using Hybrid Unstructured Meshes

Large Eddy Simulation shows considerable promise as a more accurate turbulent flow prediction tool than RANS. For typical engineering problems, the geometrical complexity means that an unstructured mesh approach is preferable. The aim of this work is to consider the dierent treatments of the convective flux term, in an unstructured formulation, and how this aects the solution accuracy. The embedded LES, or MILES approach, relies upon the numerical dissipation to remove energy from the smallest resolved scales and so comparisons are shown between embedded LES and LES using a Sub-Grid Scale model. The median dual is constructed from a mixed element mesh to provide the control volumes for the finite volume integration of the Navier-Stokes equations. Two MUSCL type flux schemes are considered. The first uses linear reconstruction, coupled with a limiter, to compute the left and right states and the dissipation term is evaluated using Roe Flux Dierence Splitting. The dissipation term is scaled using a sensor based on vorticity and divergence in order to keep the dissipative terms small. The second is computationally simpler and does not use reconstruction. The dissipative term is computed using a pseudo-Laplacian so as to give a fourth order term. Again, the magnitude of this term is controlled by a sensor. Results are shown for two cases: a fully developed pipe flow with a bulk Reynolds number of 10,000, and free jet with a Reynolds number of 36,000 and a Mach number of 0.9. All calculations produce realistic turbulent structures with no sign of re-laminarisation due to excessive numerical dissipation. For the pipe flow, the scheme using reconstruction and the full Roe FDS with MILES is observed to give weaker large scale structures, and the solution is improved considerably by scaling down the numerical dissipation and incorporating a sub-grid scale model. Good agreement is found with experiment for all computations of the free jet case, although there is some sensitivity of the potential core length to numerical scheme and presence of sub-grid scale model.