Physics-informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data

By Jian-Xun Wang, Jin-Long Wu, Heng Xiao. 2016

668 citations recorded by the source corpus.

We show that the discrepancies in Reynolds-averaged Navier-Stokes (RANS) modeled Reynolds stresses can be explained by mean flow features. A physics-informed machine learning framework is proposed to improve the predictive capabilities of RANS models by leveraging existing direct numerical simulations databases.

Subfields: 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 Computational Mechanics

In Statistical and Nonlinear Physics

Citations in the corpus

Theory-Guided Data Science: A New Paradigm for Scientific Discovery from Data (2016, background): “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-”
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 numerical data: by automating FE models to generate large numerical datasets, surrogate ML models have been trained in many applications including turbulence modeling (Wang et al., 2017), simulation of manufacturing processes and failure in advanced composit”
NSFnets (Navier-Stokes flow nets): Physics-informed neural networks for the incompressible Navier-Stokes equations (2020, methodology): “Similarly, Wang et al. [2] used random forest regression to predict the discrepancies of the baseline RANS-predicted Reynolds stresses compared to those from the DNS data, hence predicting the Reynolds stresses with high accuracy.”
Geometric Deep Learning for Volumetric Computational Fluid Dynamics (2019, background): “2 Fluid side From the fluid side, the main efforts are made to improve existing turbulent model, or generate new ones, to help getting closer to the real Navier-Stokes equations [39] [40] [41] [42] (modeling of near wall turbulent field).”
Quantification of model uncertainty in RANS simulations: A review (2018, methodology): “A machine-learning based approach to predict discrepancies of RANS-modeled Reynolds stresses has been investigated [122,123](see Section 4.”
PhyGeoNet: Physics-informed geometry-adaptive convolutional neural networks for solving parameterized steady-state PDEs on irregular domain (2020, background): “Physics-informed fully-connected neural networks Generally, training of a DNN model requires a vast amount of labeled data, which are often unavailable in many scientific machine learning (SciML) applications [9–12].”
Enhancing computational fluid dynamics with machine learning (2021, methodology): “Other ML-based approaches [50, 51] rely on physics-informed random forests to improve RANS models, with applications to cases with secondary ﬂows and separation.”
Integrating Scientific Knowledge with Machine Learning for Engineering and Environmental Systems (2020, neutral): “1) [266] [10] [219] [158] [140] [182] [203] [76] [86] [130] [174] [204] [231] [313] [121] [125] [187] [92] [306] [180] [298] [184] [267] [39] [247] [133] [130] [231] [298] [173] [66] [24] [255] [12] [17] [203] [204] [213] [239] [125] [281] [214] [176] [249] [310] [309] [311] [245] [78] [266] [297] [”
Turbulence Modeling in the Age of Data (2018, methodology): “4 Duraisamy, Iaccarino, Xiao local ow variables. While this is an activity at the L3 level (specic to a baseline model), the methodology is generally applicable to any turbulence model. In contrast, Wang et al. (2017) and Wu et al. (2018b) developed a more comprehensive perturbation strategy to pred”
Digital Twin: Values, Challenges and Enablers From a Modeling Perspective (2019, neutral): “[360] J.”