Search NASA⌕ Search

Engineering topics

Lund, T. S.

Publications and source records attributed to Lund, T. S..

16 records
All sourcesNASA NTRSDOE OSTI
NASA NTRS2014

Direct Numerical Simulation of a Weakly Stratified Turbulent Wake

Direct numerical simulation (DNS) is used to investigate a time-dependent turbulent wake evolving in a stably stratified background. A large initial Froude number is chosen to allow the wake to become fully turbulent and axisymmetric before stratification affects the spreading rate of the mean defect. The uncertainty introduced by the finite sample size associated with gathering statistics from a simulation of a time-dependent flow is reduced, compared to earlier simulations of this flow. The DNS reveals the buoyancy-induced changes to the turbulence structure, as well as to the mean-defect history and the terms in the mean-momentum and turbulence-kinetic-energy budgets, that characterize the various states of this flow - namely the three-dimensional (essentially unstratified), non-equilibrium (or 'wake-collapse') and quasi-two-dimensional (or 'two-component') regimes observed elsewhere for wakes embedded in both weakly and strongly stratified backgrounds. The wake-collapse regime is not accompanied by transfer (or 'reconversion') of the potential energy of the turbulence to the kinetic energy of the turbulence, implying that this is not an essential feature of stratified-wake dynamics. The dependence upon Reynolds number of the duration of the wake-collapse period is demonstrated, and the effect of the details of the initial/near-field conditions of the wake on its subsequent development is examined.

Redford, J. A.↗
NASA NTRS1996

Dynamic Models for LES of Turbulent Front Propagation With a Spectral Method

Direct numerical simulation of turbulent reacting flows places extreme demands on computational resources. At the present time, simulations can be performed only for greatly simplified reaction systems and for very low Reynolds numbers. Direct simulation of more realistic cases occurring at higher Reynolds number and including multiple species and numerous chemical reactions will exceed available computational resources far into the future. Because of this, there is a clear need to develop the technique of large eddy simulation for reacting flows. Unfortunately this task is complicated by the fact that combustion arises from chemical reactions that occur at the smallest scales of the flow. Capturing the large-scale behavior without resolving the small-scale details is extremely difficult in combustion problems. Thus LES modeling for turbulent combustion encounters difficulties not present in modeling momentum transport, in which the main effect of the small scales is to provide dissipation. The difficulty is more pronounced in premixed combustion, where detailed chemistry plays an essential role in determining the flame speed (or overall burning rate); in nonpremixed combustion infinite rate chemistry can be assumed, eliminating the small scale features to a first approximation.

Im, H. G.↗
NASA NTRS1996

On the Generation of Turbulent Inflow Conditions for Boundary Layer Simulations

Turbulent flows that exhibit inhomogeneities in the streamwise direction pose a particular challenge to numerical simulation approaches due to the need to prescribe time-dependent turbulent inflow conditions. In most cases the flow downstream is more or less 'driven' by the conditions at the inlet, making it necessary to specify realistic turbulent fluctuations that are in equilibrium with the assumed mean flow. This requirement often dictates that the inflow data should satisfy the Navier-Stokes equations, which in turn implies that an independent simulation be used to generate them. Detailed simulations for the purpose of creating inflow conditions can be costly and thus certain levels of approximation are desirable. In this paper we shall focus on an approximate yet accurate method for generating inflow conditions for spatially-developing boundary layer simulations. The proposed method is essentially a simplification of the method of Spalart and Leonard (1985), who devised an ingenious transformation that allows for the calculation of spatially evolving boundary layers in conjunction with periodic boundary conditions applied in the streamwise direction. While this method is elegant and highly accurate, it is more complicated than is necessary for the purpose of generating inflow data. A few key approximations are used in this work to arrive at a 'modified Spalart method' that is very easy to implement and efficient to use. The new method is shown to yield results that compare well with the computations of Spalart (1988). When used as a means of generating inflow data, the modified Spalart method is shown to be superior to existing approaches.

Lund, T. S.↗
NASA NTRS1996

Dynamic model with scale-dependent coefficients in the viscous range

The standard dynamic procedure is based on the scale-invariance assumption that the model coefficient C is the same at the grid and test-filter levels. In many applications this condition is not met, e.g. when the filter-length, delta, approaches the Kolmogorov scale, and C(delta approaches eta) approaches O. Using a priori tests, we show that the standard dynamic model yields the coefficient corresponding to the test-filter scale (alpha delta) instead of the grid-scale (delta). Several approaches to account for scale dependence are examined and/or tested in large eddy simulation of isotropic turbulence: (a) take the limit alpha approaches 1; (b) solve for two unknown coefficients C(Delta) and C(alpha delta) in the least-square-error formulation; (c) the 'bi-dynamic model', in which two test-filters (e.g. at scales 2(delta) and 4(delta) are employed to gain additional information on possible scale-dependence of the coefficient, and an improved estimate for the grid-level coefficient is obtained by extrapolation, (d) use theoretical predictions for the ratio C(alpha delta)/C(delta) and dynamically solve for C(delta). None of these options is found to be entirely satisfactory, although the last approach appears applicable to the viscous range.

Meneveau, C.↗
NASA NTRS1996

A priori testing of subgrid-scale models for the velocity-pressure and vorticity-velocity formulations

Subgrid-scale models for Large Eddy Simulation (LES) in both the velocity-pressure and the vorticity-velocity formulations were evaluated and compared in a priori tests using spectral Direct Numerical Simulation (DNS) databases of isotropic turbulence: 128(exp 3) DNS of forced turbulence (Re(sub(lambda))=95.8) filtered, using the sharp cutoff filter, to both 32(exp 3) and 16(exp 3) synthetic LES fields; 512(exp 3) DNS of decaying turbulence (Re(sub(Lambda))=63.5) filtered to both 64(exp 3) and 32(exp 3) LES fields. Gaussian and top-hat filters were also used with the 128(exp 3) database. Different LES models were evaluated for each formulation: eddy-viscosity models, hyper eddy-viscosity models, mixed models, and scale-similarity models. Correlations between exact versus modeled subgrid-scale quantities were measured at three levels: tensor (traceless), vector (solenoidal 'force'), and scalar (dissipation) levels, and for both cases of uniform and variable coefficient(s). Different choices for the 1/T scaling appearing in the eddy-viscosity were also evaluated. It was found that the models for the vorticity-velocity formulation produce higher correlations with the filtered DNS data than their counterpart in the velocity-pressure formulation. It was also found that the hyper eddy-viscosity model performs better than the eddy viscosity model, in both formulations.

Winckelmans, G. S.↗
NASA NTRS1995

Experiments with explicit filtering for LES using a finite-difference method

The equations for large-eddy simulation (LES) are derived formally by applying a spatial filter to the Navier-Stokes equations. The filter width as well as the details of the filter shape are free parameters in LES, and these can be used both to control the effective resolution of the simulation and to establish the relative importance of different portions of the resolved spectrum. An analogous, but less well justified, approach to filtering is more or less universally used in conjunction with LES using finite-difference methods. In this approach, the finite support provided by the computational mesh as well as the wavenumber-dependent truncation errors associated with the finite-difference operators are assumed to define the filter operation. This approach has the advantage that it is also 'automatic' in the sense that no explicit filtering: operations need to be performed. While it is certainly convenient to avoid the explicit filtering operation, there are some practical considerations associated with finite-difference methods that favor the use of an explicit filter. Foremost among these considerations is the issue of truncation error. All finite-difference approximations have an associated truncation error that increases with increasing wavenumber. These errors can be quite severe for the smallest resolved scales, and these errors will interfere with the dynamics of the small eddies if no corrective action is taken. Years of experience at CTR with a second-order finite-difference scheme for high Reynolds number LES has repeatedly indicated that truncation errors must be minimized in order to obtain acceptable simulation results. While the potential advantages of explicit filtering are rather clear, there is a significant cost associated with its implementation. In particular, explicit filtering reduces the effective resolution of the simulation compared with that afforded by the mesh. The resolution requirements for LES are usually set by the need to capture most of the energy-containing eddies, and if explicit filtering is used, the mesh must be enlarged so that these motions are passed by the filter. Given the high cost of explicit filtering, the following interesting question arises. Since the mesh must be expanded in order to perform the explicit filter, might it be better to take advantage of the increased resolution and simply perform an unfiltered simulation on the larger mesh? The cost of the two approaches is roughly the same, but the philosophy is rather different. In the filtered simulation, resolution is sacrificed in order to minimize the various forms of numerical error. In the unfiltered simulation, the errors are left intact, but they are concentrated at very small scales that could be dynamically unimportant from a LES perspective. Very little is known about this tradeoff and the objective of this work is to study this relationship in high Reynolds number channel flow simulations using a second-order finite-difference method.

Lund, T. S.↗
NASA NTRS1994

Passive turbulent flamelet propagation

We analyze results of a premixed constant density flame propagating in three-dimensional turbulence, where a flame model developed by Kerstein, et al. (1988) has been used. Simulations with constant and evolving velocity fields are used, where peculiar results were obtained from the constant velocity field runs. Data from the evolving flow runs with various flame speeds are used to determine two-point correlations of the fluctuating scalar field and implications for flamelet modeling are discussed.

Ashurst, William T.↗
NASA NTRS1994

A Lagrangian dynamic subgrid-scale model turbulence

A new formulation of the dynamic subgrid-scale model is tested in which the error associated with the Germano identity is minimized over flow pathlines rather than over directions of statistical homogeneity. This procedure allows the application of the dynamic model with averaging to flows in complex geometries that do not possess homogeneous directions. The characteristic Lagrangian time scale over which the averaging is performed is chosen such that the model is purely dissipative, guaranteeing numerical stability when coupled with the Smagorinsky model. The formulation is tested successfully in forced and decaying isotropic turbulence and in fully developed and transitional channel flow. In homogeneous flows, the results are similar to those of the volume-averaged dynamic model, while in channel flow, the predictions are superior to those of the plane-averaged dynamic model. The relationship between the averaged terms in the model and vortical structures (worms) that appear in the LES is investigated. Computational overhead is kept small (about 10 percent above the CPU requirements of the volume or plane-averaged dynamic model) by using an approximate scheme to advance the Lagrangian tracking through first-order Euler time integration and linear interpolation in space.

Meneveau, C.↗
NASA NTRS1992

On the local nature of the energy cascade

The local nature of the energy cascade in space and time is studied using direct numerical simulation of decaying and forced isotropic turbulence. To examine the concept that large scales evolve into smaller ones, we compute the Lagrangian correlation coefficient between local kinetic energy at different scales. This correlation is found to peak at a Lagrangian time-delay that increases with scale separation. The results show that, on average, the flow of energy to smaller scales is predominantly local in physical space and that the view of eddies decaying into smaller ones while transferring their kinetic energy appears to be, on average, quite realistic. To examine the spectral characteristics of the cascade under unsteady conditions, a pulse of large-scale energy is added to the large-eddy simulation of forced isotropic turbulence. As time progresses, the evolution of this pulse through bands of increasing wavenumbers is studied.

Meneveau, C.↗
NASA NTRS1992

Search for subgrid scale parameterization by projection pursuit regression

The dependence of subgrid-scale stresses on variables of the resolved field is studied using direct numerical simulations of isotropic turbulence, homogeneous shear flow, and channel flow. The projection pursuit algorithm, a promising new regression tool for high-dimensional data, is used to systematically search through a large collection of resolved variables, such as components of the strain rate, vorticity, velocity gradients at neighboring grid points, etc. For the case of isotropic turbulence, the search algorithm recovers the linear dependence on the rate of strain (which is necessary to transfer energy to subgrid scales) but is unable to determine any other more complex relationship. For shear flows, however, new systematic relations beyond eddy viscosity are found. For the homogeneous shear flow, the results suggest that products of the mean rotation rate tensor with both the fluctuating strain rate and fluctuating rotation rate tensors are important quantities in parameterizing the subgrid-scale stresses. A model incorporating these terms is proposed. When evaluated with direct numerical simulation data, this model significantly increases the correlation between the modeled and exact stresses, as compared with the Smagorinsky model. In the case of channel flow, the stresses are found to correlate with products of the fluctuating strain and rotation rate tensors. The mean rates of rotation or strain do not appear to be important in this case, and the model determined for homogeneous shear flow does not perform well when tested with channel flow data. Many questions remain about the physical mechanisms underlying these findings, about possible Reynolds number dependence, and, given the low level of correlations, about their impact on modeling. Nevertheless, demonstration of the existence of causal relations between sgs stresses and large-scale characteristics of turbulent shear flows, in addition to those necessary for energy transfer, provides important insight into the relation between scales in turbulent flows.

Meneveau, C.↗
NASA NTRS1987

A viscous-inviscid interaction method for ejectors having two primary jets

A viscous-inviscid interaction methodology is developed for evaluating the performance of two-dimensional incompressible flow ejectors that utilize two primary jets. In this method, the inviscid portion of the flow field, which contains the ambient fluid drawn into the device, is modeled using a higher-order panel method. The viscous region, which contains the turbulent jets, is described by an integral method. The viscous and inviscid solutions are matched together in an iterative process that simulates the interaction between the jets and the ambient fluid. The model is applied to a variety of parametric and optimization studies that illustrate the connection between the details of the ejector geometry and the level of thrust augmentation. The advantages of a dual-jet configuration are assessed through comparison with similar calculations for an otherwise identical single-jet ejector.

Lund, T. S.↗
NASA NTRS1986

Analysis of interacting dual lifting ejector systems

An analytical treatment is presented for a flowfield generated by a pair of interacting, two-dimensional parallel jets, representative of the two exhaust streams issuing from the thrust augmentor nozzles of dual lifting jet VTOL aircraft propulsion systems. Predictions of the analysis for the ratio of primary to secondary velocity are in close agreement with experimentally observed values, if the spreading rate parameter is allowed to assume a value greater than that which applies to a free jet. Theoretical results are combined with existing experimental data for unventilated jets, in order to arrive at an estimate of the thrust augmentation produced by a jet pair with an arbitrary degree of ventilation.

Lund, T. S.↗
NASA NTRS1985

An experimental investigation of the transverse pressure gradient associated with curved free jets in the proximity of solid boundaries

An experiment has been performed to determine the mass flow and thrust augmentation associated with the flow of a two-dimensional incompressible jet issued parallel to an adjacent offset wall. A theoretical analysis based on a linear static pressure profile across the jet exit reveals that thrust augmentation is possible only for a wall offset of less than one half of the nozzle exit width. Mass flow augmentation measurements are in reasonable agreement with the proposed theory. Augmented mass flow of over three percent is observed directly, while thrust augmentation in excess of three percent is predicted when measured values of the pressure difference across the exiting jet are supplied to the theory. Measurements of the separation cavity length are presented which complement those given by previous investigators by extending the range to small values of the ratio of the wall offset to the slot height. For large offset ratios it is found that the cavity length approaches 1.25 times the wall offset distance. The initial radius of curvature of the jet centerline is found to be substantially greater than the mean radius of curvature over the cavity length, thereby indicating that the trajectory of the jet does not follow the arc of a circle as has been previously assumed.

Lund, T. S.↗
NASA NTRS1985

A zonal computational procedure adapted to the optimization of two-dimensional thrust augmentor inlets

A viscous-inviscid interaction methodology based on a zonal description of the flowfield is developed as a mean of predicting the performance of two-dimensional thrust augmenting ejectors. An inviscid zone comprising the irrotational flow about the device is patched together with a viscous zone containing the turbulent mixing flow. The inviscid region is computed by a higher order panel method, while an integral method is used for the description of the viscous part. A non-linear, constrained optimization study is undertaken for the design of the inlet region. In this study, the viscous-inviscid analysis is complemented with a boundary layer calculation to account for flow separation from the walls of the inlet region. The thrust-based Reynolds number as well as the free stream velocity are shown to be important parameters in the design of a thrust augmentor inlet.

Lund, T. S.↗