Machine Learning-augmented Predictive Modeling of Turbulent Separated Flows over Airfoils

By Anand Pratap Singh, Shivaji Medida, Karthik Duraisamy. 2016

371 citations recorded by the source corpus.

A modeling paradigm is developed to augment predictive models of turbulence by effectively utilizing limited data generated from physical experiments. The key components of our approach involve inverse modeling to infer the spatial distribution of model discrepancies, and, machine learning to reconstruct discrepancy information from a large number of inverse problems into corrective model forms. We apply the methodology to turbulent flows over airfoils involving flow separation. Model augmentations are developed for the Spalart Allmaras (SA) model using adjoint-based full field inference on experimentally measured lift coefficient data. When these model forms are reconstructed using neural networks (NN) and embedded within a standard solver, we show that much improved predictions in lift can be obtained for geometries and flow conditions that were not used to train the model. The NN-augmented SA model also predicts surface pressures extremely well. Portability of this approach is demonstrated by confirming that predictive improvements are preserved when the augmentation is embedded in a different commercial finite-element solver. The broader vision is that by incorporating data that can reveal the form of the innate model discrepancy, the applicability of data-driven turbulence models can be extended to more general flows.

Subfields: Aerospace Engineering, Computational Mechanics, Statistical and Nonlinear Physics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computational Mechanics

In Statistical and Nonlinear Physics

Citations in the corpus

Theory-Guided Data Science: A New Paradigm for Scientific Discovery from Data (2016, methodology): “ples include the discovery of novel climate patterns and relationships [18], [19], closure of knowledge gaps in turbulence modeling efforts [20], [21], discovery of novel com-”
Deep learning for determining a near-optimal topological design without any iteration (2018, background): “…the potential of applying deep learning to computational physics and engineering fields, such as quantum and particle physics as well as fluid dynamics simulation (Carleo and Troyer 2017; Tompson et al. 2017; Singh et al. 2017; Mills et al. 2017) In regard to structural analysis, Lee et al .”
A Physics-Informed Machine Learning Approach for Solving Heat Transfer Equation in Advanced Manufacturing and Engineering Applications (2020, methodology): “- Theory agnostic ML models trained using physical data: Using limited physical data obtained from experiments or embedded sensors during manufacturing, predictive ML models have been developed for different applications including turbulence modeling, and guiding assembly of aerospace parts (Manohar”
Quantifying model form uncertainty in Reynolds-averaged turbulence models with Bayesian deep neural networks (2018, methodology): “1 which, in a broad sense, shares similar characteristics to earlier works on data-driven turbulence models [17, 20, 15].”
Reynolds-averaged Navier–Stokes equations with explicit data-driven Reynolds stress closure can be ill-conditioned (2018, positive): “Note, however, that some data-driven approaches, e.g., those based on correcting turbulence transport equations (Parish & Duraisamy 2016; Singh et al. 2017) or discovering analytical turbulent constitutive relations by using symbolic regression (Weatheritt & Sandberg 2016) are less affected by the…”
Quantification of model uncertainty in RANS simulations: A review (2018, methodology): “[121] showed predicted pressure coefficient of the S809 airfoil at Re 2 106 = × by using the SA model augmented by the correction function, which was trained with the inferred discrepancy field by using the data from the S814 airfoil at Reynolds numbers Re 1 106 = × and Re 2 106 = × .”
Physics-informed machine learning approach for augmenting turbulence models: A comprehensive framework (2017, methodology): “Dow and Wang [5] chose the discrepancy field ∆νt in turbulent viscosity as the response, while Duraisamy and co-workers [7, 11, 12] introduced a full-field multiplicative discrepancy factor β into the production term of the transport equation as the learning target.”
Machine learning methods for turbulence modeling in subsonic flows around airfoils (2018, methodology): “[8-11] modeled the source terms in SA turbulence model by neural networks and embedded it into CFD solver.”