A hybrid partitioned deep learning methodology for moving interface and fluid–structure interaction

Gupta, Rachit; Jaiman, Rajeev K.

doi:10.1016/j.compfluid.2021.105239

Cited by 25 publications

(8 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Typically, the high-dimensional fluid and structure dynamics are reduced into low-dimensional latent spaces using techniques like POD or convolutional auto-encoders, and the latent fluid and structure dynamics are learned by separate DNNs [23]. Physical interface constraints, such as moving interfaces (solid-to-fluid coupling) and fluid forces (fluid-to-solid coupling), are often used to couple the fluid and structure DNNs, which can be represented using methods like level-set functions [23][24][25], immersed boundary method (IBM) masks [26], or direct forcing terms [27]. By doing so, both the structural responses and the fluid dynamics can be learned in a consistent manner.…”

Section: Introductionmentioning

confidence: 99%

Differentiable hybrid neural modeling for fluid-structure interaction

Fan¹,

Wang²

2023

Preprint

View full text Add to dashboard Cite

Solving complex fluid-structure interaction (FSI) problems, which are described by nonlinear partial differential equations, is crucial in various scientific and engineering applications. Traditional computational fluid dynamics based solvers are inadequate to handle the increasing demand for large-scale and long-period simulations. The ever-increasing availability of data and rapid advancement in deep learning (DL) have opened new avenues to tackle these chal-

show abstract

Section: Introductionmentioning

confidence: 99%

Differentiable hybrid neural modeling for fluid-structure interaction

Fan¹,

Wang²

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The construction of heavily over-parametrized functions by deep neural networks rely on the foundations of the Kolmogorov-Arnold representation theorem [35] and the universal approximation of functions via neural networks [36,37]. For the data-driven modeling of nonlinear PDEs, deep neural network architectures such as convolutional recurrent autoencoder (CRAN) can be efficient and useful for constructing lowdimensional learning models [21,27,38]. CRAN is a fully data-driven approach in which both the lowdimensional representation of the state and its time evolution are learned using deep learning algorithms.…”

Section: Introductionmentioning

confidence: 99%

“…CRAN is a fully data-driven approach in which both the lowdimensional representation of the state and its time evolution are learned using deep learning algorithms. Convolutional recurrent autoencoders have been shown to perform well for unsteady flow and fluidstructure phenomenon [22,27,38]. On the other hand, the ability of current CRAN architecture to learn PDEs with a dominant hyperbolic character relies on learning low-dimensional manifold with convolutional autoencoder and evolving these low-dimensional latent representations in time via RNN-LSTM, which can pose difficulties to generalize for the various physical phenomenon characterized by hyperbolic PDEs.…”

Section: Introductionmentioning

confidence: 99%

“…On the other hand, the ability of current CRAN architecture to learn PDEs with a dominant hyperbolic character relies on learning low-dimensional manifold with convolutional autoencoder and evolving these low-dimensional latent representations in time via RNN-LSTM, which can pose difficulties to generalize for the various physical phenomenon characterized by hyperbolic PDEs. The current work build upon on our previous work on the convolutional recurrent autoencoder net for the unsteady flow dynamics and fluid-structure interaction [27,38].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning Wave Propagation with Attention-Based Convolutional Recurrent Autoencoder Net

Deo,

Jaiman

2022

Preprint

Self Cite

View full text Add to dashboard Cite

In this paper, we present an end-to-end attention-based convolutional recurrent autoencoder (AB-CRAN) network for data-driven modeling of wave propagation phenomena. The proposed network architecture relies on the attention-based recurrent neural network (RNN) with long short-term memory (LSTM) cells. To construct the low-dimensional learning model, we employ a denoising-based convolutional autoencoder from the full-order snapshots given by time-dependent hyperbolic partial differential equations for wave propagation. To begin, we attempt to address the difficulty in evolving the low-dimensional representation in time with a plain RNN-LSTM for wave propagation phenomenon. We build an attentionbased sequence-to-sequence RNN-LSTM architecture to predict the solution over a long time horizon. To demonstrate the effectiveness of the proposed learning model, we consider three benchmark problems namely one-dimensional linear convection, nonlinear viscous Burgers and two-dimensional Saint-Venant shallow water system. Using the time-series datasets from the benchmark problems, our novel AB-CRAN architecture accurately captures the wave amplitude and preserves the wave characteristics of the solution for long time horizons. The attention-based sequence-to-sequence network increases the time-horizon of prediction by five times compared to the plain RNN-LSTM. Denoising autoencoder further reduces the mean squared error of prediction and improves the generalization capability in the parameter space.

show abstract

“…The methodology alleviates the problem of knowing the underlying differential equation and the computational burden of numerical methods since neural networks with implicit bias can be efficient for inference. The CRAN methodology has been successfully applied for fluid-structure interaction 13 and underwater radiated noise 27 .…”

Section: Introductionmentioning

confidence: 99%

Combined space-time reduced-order model with 3D deep convolution for extrapolating fluid dynamics

Deo¹,

Gao²,

Jaiman³

2022

Preprint

View full text Add to dashboard Cite

There is a critical need for efficient and reliable active flow control strategies to reduce drag and noise in aerospace and marine engineering applications. While traditional full-order models based on the Navier-Stokes equations are not feasible, advanced model reduction techniques can be inefficient for active control tasks especially with strong non-linearity and convection-dominated phenomena. Using convolutional recurrent autoencoder network architectures, deep learning-based reduced-order models have been recently shown to be effective while performing several orders of magnitude faster than full-order simulations. However, these models encounter significant challenges outside the training data, limiting their effectiveness for active control and optimization tasks. In this study, we aim to improve the extrapolation capability by modifying the network architecture and integrating coupled space-time physics as an implicit bias. Reduced-order models via deep learning generally employ decoupling in spatial and temporal dimensions, which can introduce modeling and approximation errors. To alleviate these errors, we propose a novel technique for learning coupled spatial-temporal correlation using a 3D convolution network. We assess the proposed technique against a standard encoder-propagator-decoder model and demonstrate a superior extrapolation performance. To demonstrate the effectiveness of the 3D convolution network, we consider a benchmark problem of the flow past a circular cylinder at laminar flow conditions and use the spatio-temporal snapshots from the full-order simulations. Our proposed 3D convolution architecture accurately captures the velocity and pressure fields for varying Reynolds numbers. Compared to the standard encoder-propagator-decoder network, the spatio-temporal-based 3D convolution network improves the prediction range of Reynolds numbers outside of the training data.

show abstract

A hybrid partitioned deep learning methodology for moving interface and fluid–structure interaction

Cited by 25 publications

References 38 publications

Differentiable hybrid neural modeling for fluid-structure interaction

Differentiable hybrid neural modeling for fluid-structure interaction

Learning Wave Propagation with Attention-Based Convolutional Recurrent Autoencoder Net

Combined space-time reduced-order model with 3D deep convolution for extrapolating fluid dynamics

Contact Info

Product

Resources

About