SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Report discusses mathematical models of turbulence used in numerical simulations of complicated viscous, hypersonic flows. Includes survey of essential features of models and their statuses in applications.
This viewgraph presentation summarizes some CFD experience at GE Aircraft Engines for flows in the primary gaspath of a gas turbine engine and in turbine blade cooling passages. It is concluded that application of the standard k-epsilon turbulence model with wall functions is not adequate for accurate CFD simulation of aerodynamic performance and heat transfer in the primary gas path of a gas turbine engine. New models are required in the near-wall region which include more physics than wall functions. The two-layer modeling approach appears attractive because of its computational complexity. In addition, improved CFD simulation of film cooling and turbine blade internal cooling passages will require anisotropic turbulence models. New turbulence models must be practical in order to have a significant impact on the engine design process. A coordinated turbulence modeling effort between NASA centers would be beneficial to the gas turbine industry.
A statistical model of turbulence in fully developed two-dimensional incompressible turbulent mixing layers is proposed. The development of this model is largely motivated by the recent experimental observations of Brown and Roshko (1974). The model is based on the proposition that the turbulence of a fully developed two-dimensional incompressible mixing layer is in a state of quasi-equilibrium. The model is used to predict the second-order turbulence statistics of the flow including single-point turbulent Reynolds stress distribution, intensity of turbulent velocity components, rms turbulent pressure fluctuations, power spectra and two-point space-time correlation functions. It is shown that numerical results compare favorably with available experimental measurements.
A theoretical study was conducted to develop an analytical prediction method for helicopter main rotor noise due to the ingestion of atmospheric turbulence. This study incorporates an atmospheric turbulence model, a rotor mean flow contraction model and a rapid distortion turbulence model which together determine the statistics of the non-isotropic turbulence at the rotor plane. Inputs to the combined mean inflow and turbulence models are controlled by atmospheric wind characteristics and helicopter operating conditions. A generalized acoustic source model was used to predict the far field noise generated by the non-isotropic flow incident on the rotor. Absolute levels for acoustic spectra and directivity patterns were calculated for full scale helicopters, without the use of empirical or adjustable constants. Comparisons between isotropic and non-isotropic turbulence at the rotor face demonstrated pronounced differences in acoustic spectra. Turning and contraction of the flow for hover and low speed vertical ascent cases result in a 3 dB increase in the acoustic spectrum energy and a 10 dB increase in tone levels. Compared to trailing edge noise, turbulence ingestion noise is the dominant noise mechanism below approximately 30 rotor harmonics, while above 100 harmonics, trailing edge noise levels exceed turbulence ingestion noise by 25 dB.
Three dimensional combustor calculations are currently stretching the computer hardware capabilities and the computing budgets of gas turbine manufacturers. One of the main reasons for this relates to the large number of complex physical processes occurring in the combustor. Airflow, fuel spray, reaction kinetics, flame radiation, and not the least of which, turbulence must be modeled and the related differential equations solved. Discussions in this conference will address methods to improve the accuracy of combustor flow field calculations and methods to speed the convergence of the modeled equations. This report will focus on aspects of merging these two new technologies. The improved accuracy discretization schemes have a negative impact on the speed of convergence of the modeled equations that the improved solution algorithms may not overcome. A description of the causes of this problem and potential solutions will be examined.
The objectives of the Sonic Booms in Atmospheric Turbulence (SonicBAT) Program were to develop and validate, via research flight experiments under a range of realistic atmospheric conditions, one numeric turbulence model research code and one classic turbulence model research code using traditional N-wave booms in the presence of atmospheric turbulence, and to apply these models to assess the effects of turbulence on the levels of shaped sonic booms predicted from low boom aircraft designs. The SonicBAT program has successfully investigated sonic boom turbulence effects through the execution of flight experiments at two NASA centers, Armstrong Flight Research Center (AFRC) and Kennedy Space Center (KSC), collecting a comprehensive set of acoustic and atmospheric turbulence data that were used to validate the numeric and classic turbulence models developed. The validated codes were incorporated into the PCBoom sonic boom prediction software and used to estimate the effect of turbulence on the levels of shaped sonic booms associated with several low boom aircraft designs. The SonicBAT program was a four year effort that consisted of turbulence model development and refinement throughout the entire period as well as extensive flight test planning that culminated with the two research flight tests being conducted in the second and third years of the program. The SonicBAT team, led by Wyle, includes partners from the Pennsylvania State University, Lockheed Martin, Gulfstream Aerospace, Boeing, Eagle Aeronautics, Technical & Business Systems, and the Laboratory of Fluid Mechanics and Acoustics (France). A number of collaborators, including the Japan Aerospace Exploration Agency, also participated by supporting the experiments with human and equipment resources at their own expense. Three NASA centers, AFRC, Langley Research Center (LaRC), and KSC were essential to the planning and conduct of the experiments. The experiments involved precision flight of either an F-18A or F-18B executing steady, level passes at supersonic airspeeds in a turbulent atmosphere to create sonic boom signatures that had been distorted by turbulence. The flights spanned a range of atmospheric turbulence conditions at NASA Armstrong and Kennedy in order to provide a variety of conditions for code validations. The SonicBAT experiments at both sites were designed to capture simultaneous F-18A or F-18B onboard flight instrumentation data, high fidelity ground based and airborne acoustic data, surface and upper air meteorological data, and additional meteorological data from ultrasonic anemometers and SODARs to determine the local atmospheric turbulence and boundary layer height.
In this paper we study the efficacy of combining machine-learning methods with projection-based model reduction techniques for creating data-driven surrogate models of computationally expensive, high-fidelity physics models. Such surrogate models are essential for many-query applications e.g., engineering design optimization and parameter estimation, where it is necessary to invoke the high-fidelity model sequentially, many times. Surrogate models are usually constructed for individual scalar quantities. However there are scenarios where a spatially varying field needs to be modeled as a function of the model’s input parameters. Here we develop a method to do so, using projections to represent spatial variability while a machine-learned model captures the dependence of the model’s response on the inputs. The method is demonstrated on modeling the heat flux and pressure on the surface of the HIFiRE-1 geometry in a Mach 7.16 turbulent flow. The surrogate model is then used to perform Bayesian estimation of freestream conditions and parameters of the SST (Shear Stress Transport) turbulence model embedded in the high-fidelity (Reynolds-Averaged Navier–Stokes) flow simulator, using shock-tunnel data. The paper provides the first-ever Bayesian calibration of a turbulence model for complex hypersonic turbulent flows. We find that the primary issues in estimating the SST model parameters are the limited information content of the heat flux and pressure measurements and the large model-form error encountered in a certain part of the flow.
A new model for the energy transfer mechanism in the large-scale turbulent kinetic energy equation is proposed. An estimate of the characteristic length scale of the energy containing large structures is obtained from the wavelength associated with the structures predicted by a weakly nonlinear analysis for turbulent free shear flows. With the inclusion of the proposed energy transfer model, the weakly nonlinear wave models for the turbulent large-scale structures are self-contained and are likely to be independent flow geometries. The model is tested against a plane mixing layer. Reasonably good agreement is achieved. Finally, it is shown by using the Liapunov function method, the balance between the production and the drainage of the kinetic energy of the turbulent large-scale structures is asymptotically stable as their amplitude saturates. The saturation of the wave amplitude provides an alternative indicator for flow self-similarity.
The two-fluid single-column model of Thuburn et al. (Quart. J. R. Meteorol. Soc., 2019, 145, 1535–1550) is extended to include moisture and horizontal wind shear. Turbulent kinetic energy is introduced as a prognostic variable, dependence on a diagnosed boundary-layer height is removed, and subfilter fluxes are approximated using a two-fluid version of a Mellor–Yamada scheme. Three mechanisms for entrainment and detrainment processes are introduced, which represent entrainment of unstable air at the surface, forced detrainment of air at the top of the boundary/cloud layers, and turbulent mixing that relaxes the convective fluid to a reference profile. A semi-implicit Eulerian discretization replaces the semi-implicit semi-Lagrangian implementation of Thuburn et al. (Quart. J. R. Meteorol. Soc., 2019, 145, 1535–1550) to improve numerical stability and conservation. The equations for the implicit time step are solved using a quasi-Newton method, which is shown to perform well in numerical tests for conservation and convergence. The two-fluid single-column model presented in this article will be applied to simulations of shallow cumulus convection in Part III.
A novel model for turbulent droplet dispersion is formulated having significantly improved computational efficiency in comparison to the conventional point source stochastic sampling methodology. In the proposed model, a computational parcel representing a group of physical particles is considered to have a normal (Gaussian) probability density function (PDF) in three-dimensional space. The mean of each PDF is determined by Lagrangian tracking of each computational parcel, either deterministically or stochastically. The variance is represented by a turbulence-induced mean squared dispersion which is based on statistical inferences from the linearized direct modeling formulation for particle/eddy interactions. Convolution of the computational parcel PDF's produces a single PDF for the physical particle distribution profile. The validity of the new model is established by comparison with the conventional stochastic sampling method, where in each parcel is represented by a delta function distribution, for non-evaporating particles injected into simple turbulent air flows.
This paper puts forth the case for the need for improved injector design tools to meet NASA s Vision for Space Exploration goals. Requirements for this improved tool are outlined and discussed. The potential for Computational Fluid Dynamics (CFD) to meet these requirements is noted along with its current shortcomings, especially relative to demonstrated solution accuracy. The concept of verification and validation is introduced as the primary process for building and quantifying the confidence necessary for CFD to be useful as an injector design tool. The verification and validation process is considered in the context of the Marshall Space Flight Center (MSFC) Combustion Devices CFD Simulation Capability Roadmap via the Simulation Readiness Level (SRL) concept. The portion of the validation process which demonstrates the ability of a CFD code to simulate heat fluxes to a rocket engine combustor wall is the focus of the current effort. The FDNS and Loci-CHEM codes are used to simulate a shear coaxial single element G02/GH2 injector experiment. The experiment was conducted a t a chamber pressure of 750 psia using hot propellants from preburners. A measured wall temperature profile is used as a boundary condition to facilitate the calculations. Converged solutions, obtained from both codes by using wall functions with the K-E turbulence model and integrating to the wall using Mentor s baseline turbulence model, are compared to the experimental data. The initial solutions from both codes revealed significant issues with the wall function implementation associated with the recirculation zone between the shear coaxial jet and the chamber wall. The FDNS solution with a corrected implementation shows marked improvement in overall character and level of comparison to the data. With the FDNS code, integrating to the wall with Mentor s baseline turbulence model actually produce a degraded solution when compared to the wall function solution with the K--E model. The Loci-CHEM solution, produced by integrating to the wall with Mentor s baseline turbulence model, matches both the heat flux rise rate in the near injector region and the peak heat flux level very well. However, it moderately over predicts the heat fluxes downstream of the reattachment point. The Loci-CHEM solution achieved by integrating to the wall with Mentor s baseline turbulence model was clearly superior to the other solutions produced in this effort.
An improved correction has been developed to increase the accuracy with which certain formulations of computational fluid dynamics predict mixing in shear layers of hot jet flows. The CFD formulations in question are those derived from the Reynolds-averaged Navier-Stokes equations closed by means of a two-equation model of turbulence, known as the k-epsilon model, wherein effects of turbulence are summarized by means of an eddy viscosity. The need for a correction arises because it is well known among specialists in CFD that two-equation turbulence models, which were developed and calibrated for room-temperature, low Mach-number, plane-mixing-layer flows, underpredict mixing in shear layers of hot jet flows. The present correction represents an attempt to account for increased mixing that takes place in jet flows characterized by high gradients of total temperature. This correction also incorporates a commonly accepted, previously developed correction for the effect of compressibility on mixing.
Turbulence pervades most flows of engineering interest, and its prediction remains a challenge on both accuracy and cost. One promising predictive approach that reduces cost combines large-eddy simulation (LES) with simulation based on Reynolds averaged Navier–Stokes equations (RANS). This study presents a method to overcome stability and accuracy issues associated with these hybrid LES-RANS methods. The method developed involves extracting the Reynolds stresses from the upstream LES solution and then using that information to convert the downstream RANS model from a scalar eddy-viscosity model to an anisotropic nonlinear eddy-viscosity model. The method developed differs from the downstream tensorial eddy-viscosity model by being independent of the coordinate system. The method developed was evaluated by computing film cooling of a flat plate with the coolant injected through one row of circular holes. Results obtained show instabilities at the LES-to-RANS interface to be eliminated. Also, the method developed yielded solutions that compare reasonably well with those from LES, even though a significant portion of the flow is computed by the adapted anisotropic RANS model instead of LES, which significantly reduced the number of grid points and computational time needed. Since modification of the downstream RANS model is based on information extracted from the upstream LES solution, the method developed is adaptive to the problem being studied.
A study has been performed to determine best practices for generating unstructured grids for Reynolds-Averaged Navier-Stokes (RANS) simulations of jet flows. The Axisymmetric Near-Sonic Jet Case from the Turbulence Modeling Resource was used for this study: a Mach 0.985 flow through the 2-in. diameter Acoustic Reference Nozzle (ARN2). Simulations were run with FUN3D and used the Menter Shear Stress Transport (SST-V), Spalart-Allmaras (S-A), and k-kL turbulence models. The axial velocity and turbulent kinetic energy fields in the jet plume of the unstructured grid solutions were compared to those of the baseline structured grid solution provided by the Turbulence Modeling Resource. Only solutions using grids with structured-like elements in the jet plume showed good agreement with the baseline structured grid solution. Using the SST-V turbulence model, the fully unstructured grid solutions predicted the jet potential core to decay upstream of the baseline solution. With the S-A turbulence model, the unstructured grid solutions predicted the jet potential core to decay upstream of the baseline solution. The solutions using the k-kL turbulence model seemed less sensitive to grid topology. Nozzle massflow and thrust performance were also compared for all simulations. Based on the results of this study, it is currently recommended that structured-like grid elements are used in the plumes of jet flows; unstructured grid elements can be used elsewhere.
Comparison of the Jones-Launder, Ng-Spalding, Saffman-Wilcox, and Wilcox-Traci two-equation turbulence models has been conducted. It was shown that the Saffman-Wilcox and Wilcox-Traci dissipation-rate formulations admit straightforward integration through the viscous sublayer, whereas integration through the viscous sublayer is a more difficult issue with the Jones-Launder dissipation-function and the Ng-Spalding length-scale formulations. Numerical computations were conducted in which the models were applied to four equilibrium boundary layer flows including adverse, zero, and favorable pressure gradients. Computations of zero pressure gradient flow over a convex wall composed the final part of the comparison.
The flowfield induced by a single, subsonic jet exhausting perpendicularly from a flat plate into a subsonic crossflow has been numerically investigated. The test case was chosen to match available experimental data where the jet Mach number was 0.78, and the freestream Mach number was 0.13. Time-averaged solutions were obtained using the thin-layer Navier-Stokes equations and two overlapping grids. The solutions were sensitive to the radial grid clustering near the edge of the jet and to the far-field boundary conditions. Experimental data comparisons were required to determine the most appropriate jet grid and satisfactory boundary conditions. Globally, the solutions converged in about 6000 iterations. The computational results accurately showed the deflected jet and associated contrarotating vortices. The fine clustered grid in the region upstream of the jet exit allowed the horseshoe vortex in the boundary layer near the jet exit to be captured. Most importantly for aircraft applications, the computed plate pressure distributions compared favorably with the experimental data over most of the surface. However, in the wake region immediately downstream of the jet exit, where there is extensive flow separation, some discrepancies with experimental data were observed. Two turbulence models were used in this study: (1) the zero-equation, two layer Baldwin-Lomax turbulence model; and (2) one-equation Baldwin-Barth turbulence model. The turbulence models gave results which generally compared no better with experimental data than the laminar computation results.
A grid convergence study is performed to establish benchmark solutions for turbulent flows in three dimensions (3D) in support of turbulence-model verification campaign at the Turbulence Modeling Resource (TMR) website. The three benchmark cases are subsonic flows around a 3D bump and a hemisphere-cylinder configuration and a supersonic internal flow through a square duct. Reference solutions are computed for Reynolds Averaged Navier Stokes equations with the Spalart-Allmaras turbulence model using a linear eddy-viscosity model for the external flows and a nonlinear eddy-viscosity model based on a quadratic constitutive relation for the internal flow. The study involves three widely-used practical computational fluid dynamics codes developed and supported at NASA Langley Research Center: FUN3D, USM3D, and CFL3D. Reference steady-state solutions computed with these three codes on families of consistently refined grids are presented. Grid-to-grid and code-to-code variations are described in detail.
Propagation of sonic booms through turbulence reduces mean sonic boom perception metric levels and also causes considerable variability. NASA’s PCBoom suite of sonic boom acoustic propagation modules includes an approximate method for accounting for the effects of turbulence on traditional N-wave sonic booms. The current implementation is ineffective for shaped sonic booms or low-booms, and it also has limited values for turbulence and ambient input parameters. NASA’s future X-59 low-boom community noise surveys require an accurate estimate of the effects of turbulence in regions across the USA, so the module must be improved. This work presents the methods of selecting which ambient and turbulence parameters should be included in an improved PCBoom turbulence module. Turbulence and ambient data were collected from two atmospheric model databases, the Climate Forecast System Version 2 and European Centre for Medium-Range Weather Forecast Reanalysis Version 5 (ERA5), hourly from 7 AM to 7 PM local time for 10 years at 19 locations across the USA. A fully-factorial propagation analysis using these parameters would be exceedingly computationally expensive. Instead, a central composite design was chosen resulting in 45 combinations of ambient and turbulence parameters. These 45 cases effectively sample the space balancing computational burden.