A Priori Assessment of Prediction Confidence for Data-Driven Turbulence Modeling

Although Reynolds-Averaged Navier–Stokes (RANS) equations are still the dominant tool for engineering design and analysis applications involving turbulent flows, standard RANS models are known to be unreliable in many flows of engineering relevance, including flows with separation, strong pressure gradients or mean flow curvature. With increasing amounts of 3-dimensional experimental data and high fidelity simulation data from Large Eddy Simulation (LES) and Direct Numerical Simulation (DNS), data-driven turbulence modeling has become a promising approach to increase the predictive capability of RANS simulations. However, the prediction performance of data-driven models inevitably depends on the choices of training flows. This work aims to identify a quantitative measure for a priori estimation of prediction confidence in data-driven turbulence modeling. This measure represents the distance in feature space between the training flows and the flow to be predicted. Specifically, the Mahalanobis distance and the kernel density estimation (KDE) technique are used as metrics to quantify the distance between flow data sets in feature space. To examine the relationship between these two extrapolation metrics and the machine learning model prediction performance, the flow over periodic hills at Re = 10595 is used as test set and seven flows with different configurations are individually used as training sets. The results show that the prediction error of the Reynolds stress anisotropy is positively correlated with Mahalanobis distance and KDE distance, demonstrating that both extrapolation metrics can be used to estimate the prediction confidence a priori. A quantitative comparison using correlation coefficients shows that the Mahalanobis distance is less accurate in estimating the prediction confidence than KDE distance. The extrapolation metrics introduced in this work and the corresponding analysis provide an approach to aid in the choice of data source and to assess the prediction performance for data-driven turbulence modeling.

[1]  Michael Manhart,et al.  Flow over periodic hills – Numerical and experimental study in a wide range of Reynolds numbers , 2009 .

[2]  Gianluca Iaccarino,et al.  A framework for epistemic uncertainty quantification of turbulent scalar flux models for Reynolds-averaged Navier-Stokes simulations , 2013 .

[3]  Hrvoje Jasak,et al.  A tensorial approach to computational continuum mechanics using object-oriented techniques , 1998 .

[4]  Brendan D. Tracey,et al.  Application of supervised learning to quantify uncertainties in turbulence and combustion modeling , 2013 .

[5]  Trevor Hastie,et al.  The Elements of Statistical Learning , 2001 .

[6]  Michael A. Leschziner,et al.  Large-eddy simulation of turbulent boundary layer separation from a rounded step , 2012 .

[7]  G. G. Stokes.  "J." , 1890, The New Yale Book of Quotations.

[8]  P. Mahalanobis.  On the generalized distance in statistics , 1936 .

[9]  Gianluca Iaccarino,et al.  Modeling of structural uncertainties in Reynolds-averaged Navier-Stokes closures , 2013 .

[10]  J. Templeton.  Evaluation of machine learning algorithms for prediction of regions of high Reynolds averaged Navier Stokes uncertainty , 2015 .

[11]  M. Kopera.  Direct numerical simulation of turbulent flow over a backward-facing step , 2011 .

[12]  Julia Ling,et al.  Machine learning strategies for systems with invariance properties , 2016, J. Comput. Phys..

[13]  S. Murman,et al.  On the accuracy of RANS simulations with DNS data. , 2016, Physics of fluids.

[14]  Jean-Philippe Laval,et al.  Direct Numerical Simulations of Converging–Diverging Channel Flow , 2011 .

[15]  Roney L. Thompson,et al.  A methodology to evaluate statistical errors in DNS data of plane channel flows , 2016 .

[16]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[17]  Gianluca Iaccarino,et al.  Modeling Structural Uncertainties in Reynolds-Averaged Computations of Shock/Boundary Layer Interactions , 2011 .

[18]  Ulrich Schumann,et al.  Direct Numerical Simulation of Separated Turbulent Flow over a Wavy Boundary , 1996 .

[19]  Guilhem Lacaze,et al.  UNCERTAINTY ANALYSIS AND DATA-DRIVEN MODEL ADVANCES FOR A JET-IN-CROSSFLOW , 2016 .

[20]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[21]  B. Launder,et al.  Application of the energy-dissipation model of turbulence to the calculation of flow near a spinning disc , 1974 .

[22]  Julia Ling,et al.  Physics-informed machine learning approach for augmenting turbulence models: A comprehensive framework , 2017, Physical Review Fluids.

[23]  B. Launder,et al.  Development and application of a cubic eddy-viscosity model of turbulence , 1996 .

[24]  Eric R. Ziegel,et al.  The Elements of Statistical Learning , 2003, Technometrics.

[25]  F. Durst,et al.  Presentation of anisotropy properties of turbulence, invariants versus eigenvalue approaches , 2007 .

[26]  P. J. Green,et al.  Density Estimation for Statistics and Data Analysis , 1987 .

[27]  Heng Xiao,et al.  Quantifying and reducing model-form uncertainties in Reynolds-averaged Navier-Stokes simulations: A data-driven, physics-informed Bayesian approach , 2015, J. Comput. Phys..

[28]  Gianluca Iaccarino,et al.  RANS modeling of turbulent mixing for a jet in supersonic cross flow: model evaluation and uncertainty quantification , 2012 .

[29]  Anand Pratap Singh,et al.  New Approaches in Turbulence and Transition Modeling Using Data-driven Techniques , 2015 .

[30]  Peter E. Hart,et al.  Nearest neighbor pattern classification , 1967, IEEE Trans. Inf. Theory.

[31]  D. W. Scott,et al.  Multivariate Density Estimation, Theory, Practice and Visualization , 1992 .

[32]  Heng Xiao,et al.  Physics-Informed Machine Learning for Predictive Turbulence Modeling: Towards a Complete Framework. , 2016 .

[33]  Jinlong Wu,et al.  Physics-informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data , 2016, 1606.07987.

[34]  J. Templeton,et al.  Reynolds averaged turbulence modelling using deep neural networks with embedded invariance , 2016, Journal of Fluid Mechanics.

[35]  Michele Milano,et al.  Neural network modeling for near wall turbulent flow , 2002 .

[36]  Marcel Lesieur,et al.  Direct numerical simulations of high velocity ratio coaxial jets: mixing properties and influence of upstream conditions , 2007 .