Mean flow reconstruction of unsteady flows using physics-informed neural networks

Sliwinski, Lukasz; Rigas, Georgios

doi:10.1017/dce.2022.37

Cited by 5 publications

(2 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several studies have since applied PINNs for a range of scientific problems (Shukla et al, 2022), involving stochastic PDEs, integro-differential equations, fractional PDEs, non-linear differential equations (Uddin et al, 2023), and optimal control of PDEs (Mowlavi and Nabi, 2023). Within engineering, PINNs have been applied to problems in fluid dynamics (Mao et al, 2020; Raissi et al, 2020; Sliwinski and Rigas, 2023), heat transfer (Zobeiry and Humfeld, 2021), tensile membranes (Kabasi et al, 2023), and material behavior modeling (Zheng et al, 2022). However, there are only a few applications to SHM, the subject area of focus in this paper.…”

Section: Introductionmentioning

confidence: 99%

Physics-informed neural networks for structural health monitoring: a case study for Kirchhoff–Love plates

Al-Adly,

Kripakaran

2024

DCE

View full text Add to dashboard Cite

Physics-informed neural networks (PINNs), which are a recent development and incorporate physics-based knowledge into neural networks (NNs) in the form of constraints (e.g., displacement and force boundary conditions, and governing equations) or loss function, offer promise for generating digital twins of physical systems and processes. Although recent advances in PINNs have begun to address the challenges of structural health monitoring, significant issues remain unresolved, particularly in modeling the governing physics through partial differential equations (PDEs) under temporally variable loading. This paper investigates potential solutions to these challenges. Specifically, the paper will examine the performance of PINNs enforcing boundary conditions and utilizing sensor data from a limited number of locations within it, demonstrated through three case studies. Case Study 1 assumes a constant uniformly distributed load (UDL) and analyzes several setups of PINNs for four distinct simulated measurement cases obtained from a finite element model. In Case Study 2, the UDL is included as an input variable for the NNs. Results from these two case studies show that the modeling of the structure’s boundary conditions enables the PINNs to approximate the behavior of the structure without requiring satisfaction of the PDEs across the whole domain of the plate. In Case Study (3), we explore the efficacy of PINNs in a setting resembling real-world conditions, wherein the simulated measurment data incorporate deviations from idealized boundary conditions and contain measurement noise. Results illustrate that PINNs can effectively capture the overall physics of the system while managing deviations from idealized assumptions and data noise.

show abstract

Section: Introductionmentioning

confidence: 99%

Physics-informed neural networks for structural health monitoring: a case study for Kirchhoff–Love plates

Al-Adly,

Kripakaran

2024

DCE

View full text Add to dashboard Cite

show abstract

“…[13,14], to identify a spatially resolved coefficient in the k-ω and Spalart-Allmaras model, respectively, to match the RANS solutions to high-fidelity mean fields. In recent studies, closure fields are identified by assimilating RANS equations to LES, PIV [15], and DNS [16] mean fields using physics-informed neural networks.…”

Section: Introductionmentioning

confidence: 99%

Deep Reinforcement Learning-Augmented Spalart–Allmaras Turbulence Model: Application to a Turbulent Round Jet Flow

Fuchs,

von Saldern,

Kaiser

et al. 2024

Fluids

View full text Add to dashboard Cite

The purpose of this work is to explore the potential of deep reinforcement learning (DRL) as a black-box optimizer for turbulence model identification. For this, we consider a Reynolds-averaged Navier–Stokes (RANS) closure model of a round turbulent jet flow at a Reynolds number of 10,000. For this purpose, we augment the widely utilized Spalart–Allmaras turbulence model by introducing a source term that is identified by DRL. The algorithm is trained to maximize the alignment of the augmented RANS model velocity fields and time-averaged large eddy simulation (LES) reference data. It is shown that the alignment between the reference data and the results of the RANS simulation is improved by 48% using the Spalart–Allmaras model augmented with DRL compared to the standard model. The velocity field, jet spreading rate, and axial velocity decay exhibit substantially improved agreement with both the LES reference and literature data. In addition, we applied the trained model to a jet flow with a Reynolds number of 15,000, which improved the mean field alignment by 35%, demonstrating that the framework is applicable to unseen data of the same configuration at a higher Reynolds number. Overall, this work demonstrates that DRL is a promising method for RANS closure model identification. Hurdles and challenges associated with the presented methodology, such as high numerical cost, numerical stability, and sensitivity of hyperparameters are discussed in the study.

show abstract

Physics-informed deep-learning applications to experimental fluid mechanics

Eivazi,

Wang,

Vinuesa

2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

High-resolution reconstruction of flow-field data from low-resolution and noisymeasurements is of interest due to the prevalence of such problems in experimentalfluid mechanics, where the measurement data are in general sparse, incomplete andnoisy. Deep-learning approaches have been shown suitable for such super-resolutiontasks. However, a high number of high-resolution examples is needed, which may notbe available for many cases. Moreover, the obtained predictions may lack in complyingwith the physical principles, e.g. mass and momentum conservation. Physics-informeddeep learning provides frameworks for integrating data and physical laws for learning.In this study, we apply physics-informed neural networks (PINNs) for super-resolutionof flow-field data both in time and space from a limited set of noisy measurementswithout having any high-resolution reference data. Our objective is to obtain acontinuous solution of the problem, providing a physically-consistent prediction atany point in the solution domain. We demonstrate the applicability of PINNs for thesuper-resolution of flow-field data in time and space through three canonical cases:Burgers’ equation, two-dimensional vortex shedding behind a circular cylinder and theminimal turbulent channel flow. The robustness of the models is also investigatedby adding synthetic Gaussian noise. Our results show the adequate capabilities ofPINNs in the context of data augmentation for experiments in fluid mechanics.

show abstract

Mean flow reconstruction of unsteady flows using physics-informed neural networks

Cited by 5 publications

References 23 publications

Physics-informed neural networks for structural health monitoring: a case study for Kirchhoff–Love plates

Physics-informed neural networks for structural health monitoring: a case study for Kirchhoff–Love plates

Deep Reinforcement Learning-Augmented Spalart–Allmaras Turbulence Model: Application to a Turbulent Round Jet Flow

Physics-informed deep-learning applications to experimental fluid mechanics

Contact Info

Product

Resources

About