Peter Chang

Large Eddy Simulation of a Circulation Control Airfoil

A detailed study of the compressible, turbulent flow around a circulation control airfoil has been conducted using a large eddy simulation (LES). The circulation control airfoil investigated is the NCCR 1510-7067N, which has a circular arc shaped trailing edge. The Reynolds number for the problem is 5.45x10, based on chord, with a free stream Mach number of 0.12. The non-dimensional blowing rate (Cμ) for the Coanda jet is 0.093. The filtered Navier-Stokes equations are discretized using a fifth-order spatially accurate, upwind scheme with no limiting. The sub-grid scale turbulence model combines a transport equation for the sub-grid scale turbulent kinetic energy with a turbulent length scale, based on local grid dimensions, to form a expression for the sub-grid scale eddy viscosity. The resulting mean surface pressure distribution compares fairly well to experimentally measured values for the airfoil. Resulting mean velocity and Reynolds stress profiles at selected locations around the airfoil are examined.

Numerical Simulation of Two- and Three-Dimensional Circulation Control Problems

The flows about 2-D and 3-D bluff trailing edge circulation control (CC) airfoils are computed using steady Reynolds Averaged Navier-Stokes (RANS) methods. The 2-D foil is the NCCR 1510-7067 elliptical CC airfoil with circular and logarithmic spiral trailing edge geometries. The free stream Reynolds number, based on chord, is 5.45 × 10 5 , with a free stream Mach number of 0.12. For the circular trailing edge the slot height, blowing rate and angle of attack are varied, while for the logarithmic spiral only the blowing rate is varied. The 3-D foil is a semi-span wing with an elliptical cross section. It is run with a chord-based Reynolds number of 2 × 10 6 and two blowing rates. The 2-D flows are computed using the compressible, segregated solver, Fluent. 2-D results show that the full-Reynolds stress turbulence model (FRSM) predicts the correct jet detachment behavior for the circular trailing edge although the integrated lift forces are consistently underpredicted. The coanda jet detachment point for the logarithmic spiral trailing edge is predicted correctly for a lower blowing rate, but as blowing rate increases, the jet does not detach until it has wrapped around to the pressure side. We show additional 2-D results using mesh refinement via grid adaption and isotropic eddy viscosity turbulence models. The 3-D simulations use the incompressible segregated Fluent solver applying the k −ω SST turbulence model. Results show a slight attachment of the the coanda jet on the pressure side, but the results are generally encouraging.

Prediction of Vortex Shedding From a High Reynolds Number Airfoil Using LES With and Without Wall Model

ATTACHED, wall-bounded flows impose computational requirements on LES that increase drastically with Reynolds number. For that reason, even simple geometries, such as airfoils at small angles of attack, with spanwise uniform section shape, challenge the bounds of LES as chord-based Reynolds numbers increase much above 1 million. Of particular concern is the ability of LES to predict the occurrence, and strength of, weak vortex shedding from the airfoil trailing edge (by weak vortex shedding we mean that the acoustic vortex shedding signature may rise only a few decibels above that for the broadband turbulent boundary layer acoustic sources). Correct prediction of weak vortex shedding may depend on accurately predicting the flow over the entire airfoil that includes the attached, turbulent upstream flow, adverse pressure gradient and separated flow regions and finally, the turbulent wake. This paper compares results of two full-LES and two LES with wall-stress model for the flow about a modified NACA 0016 airfoil with a 41° trailing edge apex angle and a slightly convex pressure side. Comparisons of vortex shedding, as measured by the power spectral density (PSD) of wall pressure fluctuations (WPF) on the pressure side of the TE and the PSD of the vertical velocity fluctuations in the wake are made. The results indicate that vortex shedding predictions are dependent upon the stream-wise and spanwise grid resolution. In order to reduce the large computational times required for simulating the high-Reynolds number flows with fully-resolved LES, a wall-stress model that solves the turbulent boundary layer equations in the near-wall region is applied. Compared with the fully-resolved LES, the LES with wall-stress simulations require about 20 percent the number of grid points and require about 10 percent of the computational time. However, the LES with wall stress model results under-predict the vortex shedding peak in the wake and are not able to predict the vortex shedding signature in TE wall pressure spectra. These results indicate that near-wall turbulence structures need to be resolved in order to correctly predict the occurence and strength of vortex shedding.© 2006 ASME

Towards Prediction of 3-D Separated Flows Using an Unstructured LES Code

In this paper, we document large eddy simulations (LES) performed on a sphere and the Advanced Swimmer Delivery System (ASDS), a small blunt-ended submarine. The objective of this work is to develop a methodology for being able to compute the occurrence and strength of stern separation in order to accurately predict drag, maneuvering and structural loads, and acoustic signatures. We are using an energy-conserving large eddy simulations (LES) code called MPCUGLES that runs with hybrid structured-unstructured meshes. We document LES of flow over a sphere at Reynolds numbers 10,000 to 1.14 million with excellent comparison to experimental data. We also document a preliminary effort for LES of flow over ASDS. Even the simplest of these computations is very CPU-intensive necessitating the large amounts of CPU time available through the HPCMP Challenge Project C3U.

Fully-resolved LES of weakly separated flows

A high-accuracy large eddy simulation (LES) is applied to ows over a sphere and an underwater vehicle. Both objects have relatively weak separated stern ows that may depend upon the accurate resolution of turbulence structures in the attached ow regions upstream of the separation point. For the sphere we compute ows over a range of Reynolds numbers from subto super-critical (Re = 1 10 to Re = 1:14 10, respectively) for which we obtain decent agreement for the separation location, pressure distributions and integrated forces. Long time series data shows evidence of low-frequency shedding phenomena. We perform LES on the Advanced SEAL Delivery System (ASDS), an underwater vehicle with a rounded-rectangular cross section and stern slope that promotes weak ow separations. We compute the fully resolved ow over the ASDS for length-based Reynolds numbers 128 10, 256 10 and 512 10. We show that the mean ow elds over the attached ow region are reasonable in that the boundary layer pro les, shape factors and skin friction agree with other examples of developing turbulent boundary layer ows. The instantaneous ow elds exhibit near-wall turbulence structures with the correct length scales and dynamics as compared with the wall-bounded turbulent ow literature. The separation point moves aft and the extent of the separation region decreases markedly as the Reynolds number increases. Nomenclature As Surface area of sphere, m CD Integrated drag coe cient, Fx=(0:5 U oAs) Cf Local drag coe cient, w=(0:5 U o ) Cp Pressure coe cient, p=(0:5 U o ) D Sphere diameter, m Fx;y;z Force in streamwise, spanwise and vertical directions, N h Height of ASDS parallel middle body normalized by ‘gu H Boundary layer shape factor ‘gu Length of grid unit for ASDS, 5:08 cm p Pressure, N=m r;R Radii of ring-torus topology as described in the sphere results section. ReL Reynolds number for ASDS based on total length and free stream velocity. Rex Reynolds number for ASDS based on distance from bow and free stream velocity. Re Reynolds number for ASDS based on displacement thickness and free stream velocity. Corresponding author. Email: peter.chang@navy.mil, Computational Hydromechanics Division, Naval Surface Warfare Center | Carderock Division (NSWCCD), West Bethesda, MD 20817. Senior Member. yComputational Hydromechanics Division, NSWCCD, West Bethesda, MD 20817, non-member. zComputational Hydromechanics Division, NSWCCD, West Bethesda, MD 20817, non-member. xComputational Hydromechanics Division, NSWCCD, Member. {Professor, Dept. of Aerospace and Mechanical Engineering, University of Minnesota, Minneapolis, MN, 55455, Associate Fellow This material is declared a work of the U.S. Government and is not subject to copyright protection in the United States.2011

Large-Scale Computations of Unsteady Forces on Marine Vehicles

Despite efforts to maintain streamlined shapes for minimal resistance and noise, US Navy marine vehicles oftentimes must have bluff body geometries or must operate in off-design modes. These situations produce separated flows that are unsteady. At large scales the flow unsteadiness may be the cause of structural concerns while the entire range of fluid dynamic scales, from large to small, may be the cause of unwanted hydro-acoustic radiated noise. Large-eddy simulations (LES) are a class of fluid flow solvers that resolve the energy-containing scales of attached and separated flows, potentially providing the velocity field forcing functions for structural and hydro-acoustics analyses. LES have strict temporal and spatial grid spacing requirements and in order to resolve the energy-containing turbulence scales at practical flow velocities, the problem sizes and run times can be quite impressive. The objective of this Challenge Project is to apply LES to several problems of urgent US Navy need that have heretofore been considered too large to solve in a timely fashion. The first of these problems occurs when a marine vehicle, translating forward, must stop suddenly. This entails reversing the rotation of the propeller in a maneuver called crashback. This generates the largest forces that a propeller will undergo in its lifetime, and therefore, prediction of the forces and determination of procedures to reduce the forces, are of utmost importance. Using LES we have been able to uncover and verify the physics of force generation. With this Challenge Project, methodologies are being developed for one-way coupling between fluid flow and structural calculations that entail 300,000+ hours of CPU time. Another problem that is being tackled with the help of the Challenge grant is flow about the Advanced SEAL Delivery System (ASDS). This is a challenging geometry as it has both complex geometry and is a high-Reynolds number attached flow. Thus, the grid quality and size are important issues to overcome. In this paper, we document simulations of bare-hull structured grids with 4 and 20 million cells. Results of our wall-resolved simulation show that the attached boundary layers behave as expected and therefore, give us confidence that we can correctly predict the occurrence and strength of stern flow separations.

The Increasing Use of Visualization in Ship Hydrodynamics

Flow field visualization is an important part of the study of fluid dynamics and ship hydrodynamics. The field of computational fluid dynamics has provided an unprecedented ability to explore the hydrodynamics of marine vehicles through visualization. Many examples of this exist in the literature for steady flow field situations. However, unsteady visualization provides both challenges and opportunities to extract meaningful physical insight and information from computational simulations. This paper discusses some of these issues along with approaches being pursued to obtain adequate flow field information using remote high performance computing resources as well as concurrent visualization using local resources. A number of examples of flow field computations being pursued are discussed including: cavity flow, ship roll motions, trailing edge vortex shedding, ballast water exchange and crashback

Characteristics of Fuel Droplets Discharged From a Compensated Fuel/Ballast Tank

Fuel droplets, formed by the interaction of fuel plumes with a water/fuel interface, can be discharged during the refueling of water-filled compensated fuel/ballast tanks. Motivated by increasingly stringent environmental regulations, a study was initiated to understand the physical mechanisms involved in the formation and transport of fuel droplets by complex immiscible flows inside a model tank. In particular, optical measurements were made of the size distribution of fuel droplets in water discharged from a three-bay model of a compensated fuel/ballast tank. The volumetric fuel concentration of discharge from the tank was inferred from measurements of droplet size and number Flow visualizations inside the model were coupled to optical measurements of fuel droplets at the tank outlet to show that the presence of fuel in the discharged water was correlated to the formation of fuel plumes within the water-filled tank. The size distribution of fuel droplets at the tank exit is found to differ from the size distribution reported for the generation zone (near the fuel plumes) inside the tank. Thus, the advection of fuel droplets from the generation zone to the tank outlet is shown to affect the characteristics of discharged fuel droplets. The transport process specifically prevents large-diameter droplets from reaching the tank exit. Buoyancy tends to cause larger fuel droplets generated within the tank to rise and separate out of the flow before they can be discharged. The buoyancy time, τ b (D), relative to the characteristic advection time, τ a , of fuel droplets is a key parameter in predicting the fate of fuel droplets. The influence of buoyancy on the size distribution of discharged droplets was found to be modeled reasonably well by a Butterworth filter that depends on the ratio of timescales τ a /τ b (D). This model, which relates the size distribution of discharged droplets to generated droplets, is found to produce the correct qualitative behavior that larger fuel droplets are discharged when the fuel plumes move closer to the tank exit, i.e., for decreasing advection time τ a .

ENTRAINMENT CORRELATIONS BASED ON A FUEL/WATER STRATIFIED SHEAR FLOW

The purpose of this work was twofold: first, to develop correlations for the entrainment of small fuel droplets into water in a stratified fuel/water shear flow; second, to implement the correlations in a CFD code and validate it with experimental effluent fuel concentration data. It is assumed that the droplets act as passive scalars and are advected far from their generation regions where they may cause fuel contamination problems far down-stream. This work relied upon extensive experimental data obtained from a stably stratified shear flow: droplet number, droplet PDF, fluid fraction and velocity field data. The droplet data was expressed as a nondimensional entrainment velocity (E) for the volume flux of fuel due to small droplets. The fluid fraction and velocity fields at the interface were expressed in terms of Richardson numbers (Ri). It was found that E = Ce Ri−n where n = 1 and Ce is a constant, gives a good fit for the two experimental velocity cases. The best correlation was implemented in a computational simulation of the stably stratified shear flow, and the results show that the simulation can predict the entrainment quite well. A second simulation was performed for a flow with energetic vertical buoyant jets (“buoyant flow events”) and stably stratified shear flows with very large Richardson numbers. In this case, the simulations underpredicted effluent fuel concentrations by two orders of magnitude. Ad hoc corrections to the entrainment correlations show marked improvements.Copyright