RANS Simulations of a Channel Flow with a New Velocity/Pressure-Gradient Model

Poroseva, Svetlana V.; Colmenares, Juan D.; Murman, Scott M.

doi:10.2514/6.2015-3067

Cited by 7 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A new approach to quantifying uncertainty in DNS statistics that utilizes the self-consistency of RANS equations was developed in our group 6,7 . In this approach computed statistics from DNS are used to solve the RANS equations, and their solutions are compared back with the DNS data for velocity moments.…”

Section: Introductionmentioning

confidence: 99%

The Effect of the DNS Data Averaging Time on the Accuracy of RANS-DNS Simulations

Poroseva

Jeyapaul

Murman

et al. 2016

46th AIAA Fluid Dynamics Conference

Self Cite

View full text Add to dashboard Cite

Statistical data obtained from direct numerical simulations (DNS) are often used as reference data for validating turbulence models. Thus, accuracy of the DNS data itself is of particular importance for understanding the potential error in Reynolds-averaged Navier- Stokes (RANS) simulations. Recent studies demonstrate that when the DNS data is used to represent budget terms in the RANS equations, simulations of wall-bounded turbulent flows conducted with such equations (herein referred to as RANS-DNS simulations) produce unphysical results. The current paper analyzes the contribution that convergence of DNS statistics makes to this discrepancy. The Reynolds stresses and budget terms in the RANSequations are collected in a fully developed channel flow ( = ) at increasing sample sizes and analyzed using the RANS-DNS simulation. The results demonstrate that statistical convergence is not the only contributing factor to the spurious RANS-DNS results, and further study is required. Nomenclature U = mean flow velocity in the streamwise direction U ∞ = free stream velocity + = / U u τ P = mean flow pressure u, v, w = turbulent velocity fluctuations in streamwise, normal-to-wall, and spanwise directions u i = turbulent velocity fluctuation in the i-direction u τ = friction velocity h = half-channel width t = time t n = averaging time of DNS data δ = boundary layer thickness = boundary layer momentum thickness ν = kinematic viscosity 1 Assistant Professor, Mechanical Engineering, MSC01 1104, 1 UNM Albuquerque, NM, 87131-00011, AIAA Associate Fellow. 2 AIAA Member.

show abstract

Section: Introductionmentioning

confidence: 99%

The Effect of the DNS Data Averaging Time on the Accuracy of RANS-DNS Simulations

Poroseva

Jeyapaul

Murman

et al. 2016

46th AIAA Fluid Dynamics Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…Previous attempts to tackle model-form uncertainty include calibrations of a modified eddy viscosity field [9], the introduction of statistical error terms in the calibration procedure [2,10] or in the model transport equations [11], and the use of a stochastic counterpart of the Reynolds stress tensor [12]. Even UQ techniques for the RANS equations with no modelling, where the transport terms are represented with direct numerical simulation data, have been developed [13,14]. All these methods rely on detailed data from direct numerical simulations to infer about stochastic fields.…”

Section: Introductionmentioning

confidence: 99%

Bayesian Predictions of Reynolds-Averaged Navier–Stokes Uncertainties Using Maximum a Posteriori Estimates

et al. 2018

View full text Add to dashboard Cite

Computational Fluid Dynamics analyses of high Reynolds-number flows mostly rely on the Reynolds-Averaged Navier-Stokes equations. The associated closure models are based on multiple simplifying assumptions and involve numerous empirical closure coefficients, calibrated on a set of simple reference flows. Predicting new flows using a single closure model with nominal values for the closure coefficients may lead to biased predictions. Bayesian Model-Scenario Averaging is a statistical technique providing an optimal way to combine the predictions of several competing models calibrated on various sets of data (scenarios). The method allows to obtain a stochastic estimate of a Quantity-of-Interest in an unmeasured prediction scenario by i) propagating posterior probability distributions of the parameters obtained for multiple calibration scenarios, and ii) by computing a weighted posterior predictive distribution. While step ii) has a negligible computational cost, step i) requires a large number of samples of the solver.

show abstract

“…Further, even if individual terms in the closure equations are modeled accurately, small discrepancies can lead to poor results. For instance, Poroseva and Murman 10 observed that even if some terms in a second moment closure are extracted directly from DNS, the overall prediction was not satisfactory. It is thus the balance between every term in the model equation that is important.…”

Section: Introductionmentioning

confidence: 99%

Informing Turbulence Closures With Computational and Experimental Data

Duraisamy

2016

54th AIAA Aerospace Sciences Meeting

View full text Add to dashboard Cite

Data from high fidelity computational simulations and experimental measurements is used to quantify errors in the functional form of turbulence closure models. A Bayesian formalism is adopted to infer a full-field of discrepancy that provides otherwise inaccessible modeling information. The utility of the approach is demonstrated by applying it to a number of problems including channel flow, shock-boundary layer interactions, and flows with curvature and separation. In all these cases, the posterior model correlates well with the data. Furthermore, it is shown that even if limited data (such as surface pressures) is used, the accuracy of the inferred solution is improved over the entire computational domain. By directly addressing the connection between physical data and model discrepancies, the field inversion approach enhances the value of computational and experimental data for model improvement. Based on the results, it could be argued that some of the criticism of eddy viscosity models that is usually attributed to the isotropic stress-strain relationship is partially a result of poor functional forms within the model rather than the limiting nature of the Boussinesq assumption itself. Models should rather be assessed by evaluating the type and variability of corrections that will be required in a diverse set of problems. The information from the inversion procedure can be used by the modeler as a guiding tool to design more accurate model forms, or serve as input to machine learning algorithms to directly replace deficient modeling terms.

show abstract

RANS Simulations of a Channel Flow with a New Velocity/Pressure-Gradient Model

Cited by 7 publications

References 11 publications

The Effect of the DNS Data Averaging Time on the Accuracy of RANS-DNS Simulations

The Effect of the DNS Data Averaging Time on the Accuracy of RANS-DNS Simulations

Bayesian Predictions of Reynolds-Averaged Navier–Stokes Uncertainties Using Maximum a Posteriori Estimates

Informing Turbulence Closures With Computational and Experimental Data

Contact Info

Product

Resources

About