skip to main content


Title: Physics guided neural networks for spatio-temporal super-resolution of turbulent flows
Direct numerical simulation (DNS) of turbulent flows is computationally expensive and cannot be applied to flows with large Reynolds numbers. Low-resolution large eddy simulation (LES) is a popular alternative, but it is unable to capture all of the scales of turbulent transport accurately. Reconstructing DNS from low-resolution LES is critical for large-scale simulation in many scientific and engineering disciplines, but it poses many challenges to existing super-resolution methods due to the complexity of turbulent flows and computational cost of generating frequent LES data. We propose a physics-guided neural network for reconstructing frequent DNS from sparse LES data by enhancing its spatial resolution and temporal frequency. Our proposed method consists of a partial differential equation (PDE)-based recurrent unit for capturing underlying temporal processes and a physics-guided super-resolution model that incorporates additional physical constraints. We demonstrate the effectiveness of both components in reconstructing the Taylor-Green Vortex using sparse LES data. Moreover, we show that the proposed recurrent unit can preserve the physical characteristics of turbulent flows by leveraging the physical relationships in the Navier-Stokes equation.  more » « less
Award ID(s):
2203581
NSF-PAR ID:
10415804
Author(s) / Creator(s):
Date Published:
Journal Name:
Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence, PMLR
Page Range / eLocation ID:
118-128
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Simulating turbulence is critical for many societally important applications in aerospace engineering, environmental science, the energy industry, and biomedicine. Large eddy simulation (LES) has been widely used as an alternative to direct numerical simulation (DNS) for simulating turbulent flows due to its reduced computational cost. However, LES is unable to capture all of the scales of turbulent transport accurately. Reconstructing DNS from low-resolution LES is critical for many scientific and engineering disciplines, but it poses many challenges to existing super-resolution methods due to the spatio-temporal complexity of turbulent flows. In this work, we propose a new physics-guided neural network for reconstructing the sequential DNS from low-resolution LES data. The proposed method leverages the partial differential equation that underlies the flow dynamics in the design of spatio-temporal model architecture. A degradation-based refinement method is also developed to enforce physical constraints and further reduce the accumulated reconstruction errors over long periods. The results on two different types of turbulent flow data confirm the superiority of the proposed method in reconstructing the high-resolution DNS data and preserving the physical characteristics of flow transport. 
    more » « less
  2. Turbulence is a major source of momentum, heat, moisture, and aerosol transport in the atmosphere. Hence, it is crucial to understand and accurately characterize turbulence mechanisms in atmospheric flows. Many complex factors in the atmosphere influence the turbulence structures including stratification and background shear. However, our understanding of the interacting effects of these factors on coherent turbulence structure evolutions is still limited. In this talk, we aim to bridge this knowledge gap by using mode decomposition techniques and a wide range of large-eddy simulation (LES) data. By developing a data-driven technique, we will characterize unique features of atmospheric boundary layer (ABL) turbulence under different forcing scenarios. We will present 3D LES wind speed snapshots of different ABL flows that will be used as dynamic mode decomposition (DMD) input data. Then, the obtained modes and eigenvalues will be employed to gain insights into coherent turbulence structures in ABLs. We will explain the physical meaning of dominant modes and how each mode relates to the physical cause of turbulence structures. The dominant modes, which are selected based on the mode amplitude, contain the most important spatial and temporal characteristics of the flow. We will evaluate the accuracy of the performance of this method by reconstructing the flow field with only a small number of modes, and then calculate the mean average error between the real flow and the reconstructed flow fields. We will present different data frequencies, wind speeds, and surface heat fluxes. This allows us to elucidate the modes and determine the conditions in which the mode decomposition provides more accurate results for the ABL flows. Our findings can be used to identify the major causes of turbulence in real atmospheric flows and could provide a deeper insight into the dynamics of turbulence in ABLs. Our results will also be useful for developing reduced-order models that can rapidly predict the turbulent ABL flow fields. 
    more » « less
  3. Near-wall flow simulation remains a central challenge in aerodynamics modelling: Reynolds-averaged Navier–Stokes predictions of separated flows are often inaccurate, and large-eddy simulation (LES) can require prohibitively small near-wall mesh sizes. A deep learning (DL) closure model for LES is developed by introducing untrained neural networks into the governing equations and training in situ for incompressible flows around rectangular prisms at moderate Reynolds numbers. The DL-LES models are trained using adjoint partial differential equation (PDE) optimization methods to match, as closely as possible, direct numerical simulation (DNS) data. They are then evaluated out-of-sample – for aspect ratios, Reynolds numbers and bluff-body geometries not included in the training data – and compared with standard LES models. The DL-LES models outperform these models and are able to achieve accurate LES predictions on a relatively coarse mesh (downsampled from the DNS mesh by factors of four or eight in each Cartesian direction). We study the accuracy of the DL-LES model for predicting the drag coefficient, near-wall and far-field mean flow, and resolved Reynolds stress. A crucial challenge is that the LES quantities of interest are the steady-state flow statistics; for example, a time-averaged velocity component $\langle {u}_i\rangle (x) = \lim _{t \rightarrow \infty } ({1}/{t}) \int _0^t u_i(s,x)\, {\rm d}s$ . Calculating the steady-state flow statistics therefore requires simulating the DL-LES equations over a large number of flow times through the domain. It is a non-trivial question whether an unsteady PDE model with a functional form defined by a deep neural network can remain stable and accurate on $t \in [0, \infty )$ , especially when trained over comparatively short time intervals. Our results demonstrate that the DL-LES models are accurate and stable over long time horizons, which enables the estimation of the steady-state mean velocity, fluctuations and drag coefficient of turbulent flows around bluff bodies relevant to aerodynamics applications. 
    more » « less
  4. Turbulent boundary layers subject to severe acceleration or strong favorable pressure gradient (FPG) are of fundamental and technological importance. Scientifically, they elicit great interest from the points of view of scaling laws, the complex interaction between the outer and inner regions, and the quasi-laminarization phenomenon. Many flows of industrial and technological applications are subject to strong acceleration such as convergent ducts, turbines blades and nozzles. Our recent numerical predictions (J. Fluid Mech., vol. 775, pp. 189-200, 2015) of turbulent boundary layers subject to very strong FPG with high spatial/temporal resolution, i.e. Direct Numerical Simulation (DNS), have shown a meaningful weakening of the Reynolds shear stresses with an evident logarithmic behavior. In the present study, assessment of three different turbulence models (Shear Stress Transport, k-w and Spalart-Allmaras, henceforth SST, k-w and SA, respectively) in Reynolds-averaged Navier-Stokes (RANS) simulations is performed. The main objective is to evaluate the ability of popular turbulence models in capturing the characteristic features present during the quasi-laminarization phenomenon in highly accelerating turbulent boundary layers. Favorable pressure gradient is prescribed by a top converging surface (sink flow) with an approximately constant acceleration parameter of K = 4.0 x 10^(-6). Furthermore, the quasi-laminarization effect on the temperature field is also examined by solving the energy equation and assuming the temperature as a passive scalar. Validation of RANS results is carried out by means of a large DNS dataset. 
    more » « less
  5. Abstract

    Tissue dynamics play critical roles in many physiological functions and provide important metrics for clinical diagnosis. Capturing real-time high-resolution 3D images of tissue dynamics, however, remains a challenge. This study presents a hybrid physics-informed neural network algorithm that infers 3D flow-induced tissue dynamics and other physical quantities from sparse 2D images. The algorithm combines a recurrent neural network model of soft tissue with a differentiable fluid solver, leveraging prior knowledge in solid mechanics to project the governing equation on a discrete eigen space. The algorithm uses a Long-short-term memory-based recurrent encoder-decoder connected with a fully connected neural network to capture the temporal dependence of flow-structure-interaction. The effectiveness and merit of the proposed algorithm is demonstrated on synthetic data from a canine vocal fold model and experimental data from excised pigeon syringes. The results showed that the algorithm accurately reconstructs 3D vocal dynamics, aerodynamics, and acoustics from sparse 2D vibration profiles.

     
    more » « less