Showing 1–50 of 52 results for author: Jaiman, R

Predicting Wave Dynamics using Deep Learning with Multistep Integration Inspired Attention and Physics-Based Loss Decomposition

Authors: Indu Kant Deo, Rajeev K. Jaiman

Abstract: In this paper, we present a physics-based deep learning framework for data-driven prediction of wave propagation in fluid media. The proposed approach, termed Multistep Integration-Inspired Attention (MI2A), combines a denoising-based convolutional autoencoder for reduced latent representation with an attention-based recurrent neural network with long-short-term memory cells for time evolution of… ▽ More In this paper, we present a physics-based deep learning framework for data-driven prediction of wave propagation in fluid media. The proposed approach, termed Multistep Integration-Inspired Attention (MI2A), combines a denoising-based convolutional autoencoder for reduced latent representation with an attention-based recurrent neural network with long-short-term memory cells for time evolution of reduced coordinates. This proposed architecture draws inspiration from classical linear multistep methods to enhance stability and long-horizon accuracy in latent-time integration. Despite the efficiency of hybrid neural architectures in modeling wave dynamics, autoregressive predictions are often prone to accumulating phase and amplitude errors over time. To mitigate this issue within the MI2A framework, we introduce a novel loss decomposition strategy that explicitly separates the training loss function into distinct phase and amplitude components. We assess the performance of MI2A against two baseline reduced-order models trained with standard mean-squared error loss: a sequence-to-sequence recurrent neural network and a variant using Luong-style attention. To demonstrate the effectiveness of the MI2A model, we consider three benchmark wave propagation problems of increasing complexity, namely one-dimensional linear convection, the nonlinear viscous Burgers equation, and the two-dimensional Saint-Venant shallow water system. Our results demonstrate that the MI2A framework significantly improves the accuracy and stability of long-term predictions, accurately preserving wave amplitude and phase characteristics. Compared to the standard long-short term memory and attention-based models, MI2A-based deep learning exhibits superior generalization and temporal accuracy, making it a promising tool for real-time wave modeling. △ Less

Submitted 15 April, 2025; originally announced April 2025.

Comments: 30 pages, 14 figures

Data-driven modeling of fluid flow around rotating structures with graph neural networks

Authors: Rui Gao, Zhi Cheng, Rajeev K. Jaiman

Abstract: Graph neural networks, recently introduced into the field of fluid flow surrogate modeling, have been successfully applied to model the temporal evolution of various fluid flow systems. Existing applications, however, are mostly restricted to cases where the domain is time-invariant. The present work extends the application of graph neural network-based modeling to fluid flow around structures rot… ▽ More Graph neural networks, recently introduced into the field of fluid flow surrogate modeling, have been successfully applied to model the temporal evolution of various fluid flow systems. Existing applications, however, are mostly restricted to cases where the domain is time-invariant. The present work extends the application of graph neural network-based modeling to fluid flow around structures rotating with respect to a certain axis. Specifically, we propose to apply a graph neural network-based surrogate modeling for fluid flow with the mesh corotating with the structure. Unlike conventional data-driven approaches that rely on structured Cartesian meshes, our framework operates on unstructured co-rotating meshes, enforcing rotation equivariance of the learned model by leveraging co-rotating polar (2D) and cylindrical (3D) coordinate systems. To model the pressure for systems without Dirichlet pressure boundaries, we propose a novel local directed pressure difference formulation that is invariant to the reference pressure point and value. For flow systems with large mesh sizes, we introduce a scheme to train the network in single or distributed graphics processing units by accumulating the backpropagated gradients from partitions of the mesh. The effectiveness of our proposed framework is examined on two test cases: (i) fluid flow in a 2D rotating mixer, and (ii) the flow past a 3D rotating cube. Our results show that the model achieves stable and accurate rollouts for over 2000 time steps in periodic regimes while capturing accurate short-term dynamics in chaotic flow regimes. In addition, the drag and lift force predictions closely match the CFD calculations, highlighting the potential of the framework for modeling both periodic and chaotic fluid flow around rotating structures. △ Less

Submitted 28 March, 2025; originally announced March 2025.

A Graph Neural Network Surrogate Model for Multi-Objective Fluid-Acoustic Shape Optimization

Authors: Farnoosh Hadizadeh, Wrik Mallik, Rajeev K. Jaiman

Abstract: This article presents a graph neural network (GNN) based surrogate modeling approach for fluid-acoustic shape optimization. The GNN model transforms mesh-based simulations into a computational graph, enabling global prediction of pressure and velocity flow fields around solid boundaries. We employ signed distance functions to implicitly represent geometries on unstructured nodes represented by the… ▽ More This article presents a graph neural network (GNN) based surrogate modeling approach for fluid-acoustic shape optimization. The GNN model transforms mesh-based simulations into a computational graph, enabling global prediction of pressure and velocity flow fields around solid boundaries. We employ signed distance functions to implicitly represent geometries on unstructured nodes represented by the graph neural network. The trained graph neural network is employed here to predict the flow field around various airfoil shapes. The median relative error in the prediction of pressure and velocity for 300 test cases is 1-2\%. The predicted flow field is employed to extract the fluid force coefficients and the velocity profile of the boundary layer. The boundary layer velocity profile is then used to predict the flow field and noise levels, allowing the direct integration of the coupled fluid-acoustic analysis in the shape optimization algorithm. The fluid-acoustic shape optimization is extended to multi-objective shape optimization by minimizing trailing edge noise while maximizing the aerodynamic performance of airfoil surfaces. The results show that the overall sound pressure level of the optimized airfoil decreases by 13.9\% (15.82 dBA), and the lift coefficient increases by 7.2\%, for a fixed set of operating conditions. The proposed GNN-based integrated surrogate modeling with the shape optimization algorithm exhibits a computational speed-up of three orders of magnitude compared to while maintaining reasonable accuracy compared to full-order online optimization applications. The GNN-based surrogate model offers an efficient computational framework for fluid-acoustic shape optimization via adaptive morphing of structures. △ Less

Submitted 21 December, 2024; originally announced December 2024.

Comments: 29 pages, 17 figures

Flow-Induced Vibration of Flexible Hydrofoil Within Cavitating Turbulent Flow

Authors: Zhi Cheng, Rajeev Jaiman

Abstract: The flow-induced vibration and cavitation dynamics of three-dimensional flow past a cantilever flexible hydrofoil are investigated using a large eddy simulation (LES) model, a homogeneous mixture cavitation model and the structural modes superposition method. The present work aims to explore a potential mechanism responsible for a propeller singing behavior, and thus focuses on the synchronized hy… ▽ More The flow-induced vibration and cavitation dynamics of three-dimensional flow past a cantilever flexible hydrofoil are investigated using a large eddy simulation (LES) model, a homogeneous mixture cavitation model and the structural modes superposition method. The present work aims to explore a potential mechanism responsible for a propeller singing behavior, and thus focuses on the synchronized hydroelastic coupling among the pressure pulsation inside the flow field, the cavitation generation and the structural vibration. To begin, we validate the tip vortex dynamics of a flexible hydrofoil against the available experimental. Our results demonstrate that the tip vortex shedding and the blade vibration are responsible for the intense peak in the low-frequency tonal components of the noise source, and the trailing-edge vortex shedding induces broadband components. Additionally, the generation of sheet cavitation induces considerable synchronized hydrofoil vibration (subjected to a flutter-like response), and affects the pressure fluctuations in the flow field, which further dominate the features of the underwater noise sources. It is suggested that the cavitation behavior and structural vibrations co-dominate the characteristics of singing noise from a propeller blade. △ Less

Submitted 2 June, 2024; originally announced June 2024.

Report number: OMAE 2024-125985

Modeling of Hydroacoustic Noise from Marine Propellers with Tip Vortex Cavitation

Authors: Zhi Cheng, Suraj Kashyap, Brendan Smoker, Giorgio Burella, Rajeev Jaiman

Abstract: The present work aims to study the cavitating turbulent flow of a full-scale marine propeller and explore the physical mechanism underpinning the underwater radiated noise. We employ the standard dynamic large eddy simulation for the turbulent wake flow and the Schnerr-Sauer cavitation model, while the Ffowcs-Williams-Hawkings acoustic analogy is considered for the hydroacoustic modeling. For the… ▽ More The present work aims to study the cavitating turbulent flow of a full-scale marine propeller and explore the physical mechanism underpinning the underwater radiated noise. We employ the standard dynamic large eddy simulation for the turbulent wake flow and the Schnerr-Sauer cavitation model, while the Ffowcs-Williams-Hawkings acoustic analogy is considered for the hydroacoustic modeling. For the current investigation, we consider a well-known Potsdam Propeller Test Case to analyze the turbulent cavitating flow and the associated hydroacoustic emissions. To begin, the modeling framework is validated using the available experimental data, and distinctive double-helical tip vortex cavitation and its qualitative patterns along the vortex trajectory are captured. In comparison to the non-cavitating condition, the pressure distribution on the propeller surface is more disordered for the cavitating condition, which is further reflected by a relatively stronger power of both low-frequency tonal peaks and high-frequency broadband components in the spectrum of thrust generation. Specifically, the generation of cavitation leads to the enhancement of the monopole noise source and the breakdown of cavitation bubbles as well as vortex structures in the turbulent wake. Furthermore, the tonal noise with the frequency corresponding to the harmonics of blade passing frequency is also enhanced. Generally speaking, the generation of cavitation structures enhances the hydroacoustics energy of URN at all orientations, especially in the downstream direction with sound pressure level increasing up to 20 dB. △ Less

Submitted 23 May, 2024; originally announced May 2024.

Report number: OMAE2024-125991

An Efficient Phase-field Framework for Contact Dynamics between Deformable Solids in Fluid Flow

Authors: Biswajeet Rath, Xiaoyu Mao, Rajeev K. Jaiman

Abstract: Elastic contact in hydrodynamic environments is a complex multiphysics phenomenon and can be found in applications ranging from engineering to biological systems. Understanding the intricacies of this coupled problem requires the development of a generalized framework capable of handling topological changes and transitioning implicitly from FSI conditions to solid-solid contact conditions. We prop… ▽ More Elastic contact in hydrodynamic environments is a complex multiphysics phenomenon and can be found in applications ranging from engineering to biological systems. Understanding the intricacies of this coupled problem requires the development of a generalized framework capable of handling topological changes and transitioning implicitly from FSI conditions to solid-solid contact conditions. We propose a mono-field interface advancing method for handling multibody contact simulations in submerged environments. Given the physical demands of the problem, we adopt a phase-field based fully Eulerian approach to resolve the multiphase and multibody interactions in the system. We employ a stabilized finite element formulation and a partitioned iterative procedure to solve the unified momentum equation comprising solid and fluid dynamics coupled with the Allen-Cahn phase-field equation. We introduce a contact force approach to handle smooth elastic-elastic and elastic-rigid contact based on the overlap of the diffused interfaces of two colliding bodies. We propose a novel approach to extend the model for multibody contact simulations while using a single phase-field function for all the solids. The method is based on updating the solid boundaries at every time step and checking for collisions among them. The developed approach eliminates the need to solve multiple phase field equations and multiple strain equations at every time step. This reduces the overall computational time by nearly $16\%$ compared to a multi phase-field approach. The implemented model is verified for smooth dry contact and FSI contact scenarios. Using the proposed framework, we demonstrate the collision dynamics between multiple bodies submerged in an open liquid tank. △ Less

Submitted 15 September, 2024; v1 submitted 9 May, 2024; originally announced May 2024.

Comments: 35 pages

Continual Learning of Range-Dependent Transmission Loss for Underwater Acoustic using Conditional Convolutional Neural Net

Authors: Indu Kant Deo, Akash Venkateshwaran, Rajeev K. Jaiman

Abstract: There is a significant need for precise and reliable forecasting of the far-field noise emanating from shipping vessels. Conventional full-order models based on the Navier-Stokes equations are unsuitable, and sophisticated model reduction methods may be ineffective for accurately predicting far-field noise in environments with seamounts and significant variations in bathymetry. Recent advances in… ▽ More There is a significant need for precise and reliable forecasting of the far-field noise emanating from shipping vessels. Conventional full-order models based on the Navier-Stokes equations are unsuitable, and sophisticated model reduction methods may be ineffective for accurately predicting far-field noise in environments with seamounts and significant variations in bathymetry. Recent advances in reduced-order models, particularly those based on convolutional and recurrent neural networks, offer a faster and more accurate alternative. These models use convolutional neural networks to reduce data dimensions effectively. However, current deep-learning models face challenges in predicting wave propagation over long periods and for remote locations, often relying on auto-regressive prediction and lacking far-field bathymetry information. This research aims to improve the accuracy of deep-learning models for predicting underwater radiated noise in far-field scenarios. We propose a novel range-conditional convolutional neural network that incorporates ocean bathymetry data into the input. By integrating this architecture into a continual learning framework, we aim to generalize the model for varying bathymetry worldwide. To demonstrate the effectiveness of our approach, we analyze our model on several test cases and a benchmark scenario involving far-field prediction over Dickin's seamount in the Northeast Pacific. Our proposed architecture effectively captures transmission loss over a range-dependent, varying bathymetry profile. This architecture can be integrated into an adaptive management system for underwater radiated noise, providing real-time end-to-end mapping between near-field ship noise sources and received noise at the marine mammal's location. △ Less

Submitted 11 April, 2024; originally announced April 2024.

Comments: 14 pages, 18 figures

A 3D phase-field based Eulerian variational framework for multiphase fluid-structure interaction with contact dynamics

Authors: Xiaoyu Mao, Rajeev Jaiman

Abstract: Using a fixed Eulerian mesh, the phase-field method has been successfully utilized for a broad range of moving boundary problems involving multiphase fluids and single-phase fluid-structure interaction. Nevertheless, multiphase fluids interacting with multiple solids are rarely explored, especially for large-scale finite element simulations with contact dynamics. In this work, we introduce a novel… ▽ More Using a fixed Eulerian mesh, the phase-field method has been successfully utilized for a broad range of moving boundary problems involving multiphase fluids and single-phase fluid-structure interaction. Nevertheless, multiphase fluids interacting with multiple solids are rarely explored, especially for large-scale finite element simulations with contact dynamics. In this work, we introduce a novel parallelized three-dimensional fully Eulerian variational framework for simulating multiphase fluids interacting with multiple deformable solids subjected to contact dynamics. In the framework, each solid or fluid phase is identified by a standalone phase indicator. Moreover the phase indicators are initialized by the grid cell method, which restricts the calculation to several grid cells. A diffuse interface description is employed for a smooth interpolation of the physical properties across the phases, yielding unified mass and momentum conservation equations for the coupled dynamical interactions. For each solid object, temporal integration is carried out to track the strain evolution in an Eulerian frame of reference. The coupled differential equations are solved in a partitioned iterative manner. We first verify the framework against reference numerical data in a two-dimensional case of a rotational disk in a lid-driven cavity flow. The case is generalized to a rotational sphere in a lid-driven cavity flow to showcase large deformation and rotational motion of solids and examine the convergence in three dimensions. We then simulate the falling of an immersed solid sphere on an elastic block under gravitational force to demonstrate the translational motion and the solid-to-solid contact in a fluid environment. Finally, we demonstrate the framework for a ship-ice interaction problem involving multiphase fluids with an air-water interface and contact between a floating ship and ice floes. △ Less

Submitted 15 February, 2024; originally announced February 2024.

Comments: 32 pages, 18 figures

A finite element framework for fluid-structure interaction of turbulent cavitating flows with flexible structures

Authors: Nihar B. Darbhamulla, Rajeev K. Jaiman

Abstract: We present a finite element framework for the numerical prediction of cavitating turbulent flows interacting with flexible structures. The vapor-fluid phases are captured through a homogeneous mixture model, with a scalar transport equation governing the spatio-temporal evolution of cavitation dynamics. High-density gradients in the two-phase cavitating flow motivate the use of a positivity-preser… ▽ More We present a finite element framework for the numerical prediction of cavitating turbulent flows interacting with flexible structures. The vapor-fluid phases are captured through a homogeneous mixture model, with a scalar transport equation governing the spatio-temporal evolution of cavitation dynamics. High-density gradients in the two-phase cavitating flow motivate the use of a positivity-preserving Petrov-Galerkin stabilization method in the variational framework. A mass transfer source term introduces local compressibility effects arising as a consequence of phase change. The turbulent fluid flow is modeled through a dynamic subgrid-scale method for large eddy simulations. The flexible structure is represented by a set of eigenmodes, obtained through the modal analysis of the linear elasticity equations. A partitioned iterative approach is adopted to couple the structural dynamics and cavitating fluid flow, where the deforming flow domain is described by an arbitrary Lagrangian-Eulerian frame of reference. Through the numerical validation study, we establish the fidelity of the proposed framework by comparing it against experimental and numerical studies for both rigid and flexible hydrofoils in cavitating flows. Based on the validation study conducted over a flexible NACA66 rectangular hydrofoil, we elucidate the role of cavity and vortex shedding in governing the structural dynamics of a flexible hydrofoil. △ Less

Submitted 28 December, 2023; originally announced December 2023.

Comments: 37 pages, 18 figures, 8 tables

A numerical study on the oscillatory dynamics of tip vortex cavitation

Authors: Saman Lak, Rajeev Jaiman

Abstract: In this paper, we numerically study the oscillatory dynamics associated with the tip vortex cavitation over an elliptical hydrofoil section using our 3D variational multiphase flow solver at a Reynolds number of $Re=8.95 \times 10^5$ via dynamic subgrid-scale modeling and homogeneous mixture theory. To begin, we examine the grid resolution requirements and introduce a length scale that considers b… ▽ More In this paper, we numerically study the oscillatory dynamics associated with the tip vortex cavitation over an elliptical hydrofoil section using our 3D variational multiphase flow solver at a Reynolds number of $Re=8.95 \times 10^5$ via dynamic subgrid-scale modeling and homogeneous mixture theory. To begin, we examine the grid resolution requirements and introduce a length scale that considers both the tip vortex strength and the core radius, which is then employed to non-dimensionalize the spatial resolution in the tip vortex region and establish mesh requirements for large eddy simulation of tip vortex cavitation. We next perform simulations to analyze the dynamical modes of cavity oscillation at different cavitation numbers and compare them with the semi-analytical solution. The breathing mode of cavity surface oscillation is extracted from the results through the definition of an effective radius. The time-averaged effective radius demonstrates that the cavity experiences a growth region followed by decay as it progresses away from the tip. Further examination of the local breathing mode oscillations in these regions indicates the different behavior of cavity oscillations in the growth and decay regions, with the oscillations within the former being characterized by lower frequencies. For representative cavitation numbers $σ\in [1.2,2.6]$, we find that pressure fluctuations exhibit a shift of spectrum towards lower frequencies as the cavitation number decreases, similar to its influence on breathing mode oscillations. The results indicate the correlations between the breathing mode oscillations and the pressure fluctuations, with the low-frequency and relatively higher-frequency pressure fluctuations being correlated with the growth and decay regions, respectively. △ Less

Submitted 24 December, 2023; originally announced December 2023.

Comments: 30 pages, 21 figures

Journal ref: J. Fluid Mech. 998 (2024) A13

Wake Interference Effects on Flapping Dynamics of Elastic Inverted Foil

Authors: Aarshana R. Parekh, Rajeev K. Jaiman

Abstract: Using high-fidelity simulations, we study the self-induced flapping dynamics of an inverted elastic foil when it is placed in tandem with a stationary circular cylinder. The effect of wake interference on the inverted foil's coupled dynamics is examined at a fixed Reynolds number ($Re$) as a function of non-dimensional bending rigidity ($K_{B}$) and the structure to fluid mass ratio ($m^{*}$). Our… ▽ More Using high-fidelity simulations, we study the self-induced flapping dynamics of an inverted elastic foil when it is placed in tandem with a stationary circular cylinder. The effect of wake interference on the inverted foil's coupled dynamics is examined at a fixed Reynolds number ($Re$) as a function of non-dimensional bending rigidity ($K_{B}$) and the structure to fluid mass ratio ($m^{*}$). Our results show that there exists a critical $K_{B, Cr} = 0.25$, above which the downstream foil is synchronized with the unsteady wake, and the cylinder controls the flapping response and the wake vortex dynamics. During synchronization, two additional flapping modes namely, small and moderate amplitude flapping mode are observed as a function of decreasing $K_{B}$. Below $K_{B, Cr}$, the downstream foil undergoes self-induced large-amplitude flapping (LAF) similar to an isolated foil counterpart. When the dynamics of the downstream foil are analyzed for a range of $m^{*}$, we can characterize the response dynamics into two regions, namely low and high sensitivity. The high sensitivity region is observed when the dynamics are controlled by the cylinder oscillations, i.e., for foils with high stiffness. In this regime, the foil dynamics negatively correlate to $K_{B}$ and $m^{*}$. The low sensitivity region is observed when the downstream foil is no longer synchronized with the wake and undergoes an LAF response, with dynamics that are weakly correlated to $K_{B}$. A new non-dimensional parameter is proposed that combines the effect of the foil's inertia and elastic forces and can capture the foil's response when it is subjected to wake interference effects. The findings from this study aim to generalize our understanding of the self-induced flapping dynamics of inverted foils in an array configuration and have relevance to the development of inverted foil-based renewable energy harvesters. △ Less

Submitted 25 April, 2024; v1 submitted 15 November, 2023; originally announced November 2023.

Comments: 19 pages, 19 figures

Coupled dynamics of steady jet flow control for flexible membrane wings

Authors: Guojun Li, Rajeev Kumar Jaiman, Hongzhong Liu

Abstract: We present a steady jet flow-based flow control of flexible membrane wings for an adaptive and efficient motion of bat-inspired drones in complex flight environments. A body-fitted variational computational aeroelastic framework is adopted for the modeling of fluid-structure interactions. High-momentum jet flows are injected from the leading edge and transported to the wake flows to alter the aero… ▽ More We present a steady jet flow-based flow control of flexible membrane wings for an adaptive and efficient motion of bat-inspired drones in complex flight environments. A body-fitted variational computational aeroelastic framework is adopted for the modeling of fluid-structure interactions. High-momentum jet flows are injected from the leading edge and transported to the wake flows to alter the aerodynamic performance and the membrane vibration. The phase diagrams of the coupled fluid-membrane dynamics are constructed in the parameter space of the angle of attack and the jet momentum coefficient. The coupled dynamical effect of active jet flow control on the membrane performance is systematically explored. While the results indicate that the current active flow control strategy performs well at low angles of attack, the effectiveness degrades at high angles of attack with large flow separation. To understand the coupling mechanism, the variations of the vortex patterns at different jet momentum coefficients are examined by the proper orthogonal decomposition modes in the Eulerian view and the fluid transport process is studied by the coherent flow structures in the Lagrange description. Two scaling relations that quantitatively connect the membrane deformation with the aerodynamic loads presented in our previous work are verified even when active jet flow control is applied. A unifying feedback loop that reveals the fluid-membrane coupling mechanism is proposed. This feedback loop provides useful guidance for designing optimal active flow control strategies and enhancing flight capabilities. These findings can facilitate the development of next-generation bio-inspired drones that incorporate smart sensing and intelligent control. △ Less

Submitted 12 November, 2023; originally announced November 2023.

Flow-induced vibration of a flexible cantilever in tandem configuration

Authors: Shayan Heydari, Rajeev K Jaiman

Abstract: The present work investigates the fluid-structure interaction (FSI) of a flexible cylindrical cantilever in a tandem configuration. A fully coupled fluid-structure solver based on the three-dimensional incompressible Navier-Stokes equations and Euler-Bernoulli beam equation is employed to numerically examine the coupled dynamics of the cantilever. We assess the extent to which such a flexible stru… ▽ More The present work investigates the fluid-structure interaction (FSI) of a flexible cylindrical cantilever in a tandem configuration. A fully coupled fluid-structure solver based on the three-dimensional incompressible Navier-Stokes equations and Euler-Bernoulli beam equation is employed to numerically examine the coupled dynamics of the cantilever. We assess the extent to which such a flexible structure could sustain oscillations in both subcritical and post-critical regimes of Reynolds number ($Re$). Spatio-temporal power transfer patterns, response amplitudes, and vorticity dynamics are quantified and compared between isolated and tandem configurations. Results of our analysis indicate that the cantilever in tandem configuration is prone to sustained oscillations dependent on $Re$ and the reduced velocity parameter ($U^*$). In the subcritical $Re$ regime, the cantilever exhibits sustained oscillations with peak transverse oscillation amplitudes occurring within a specific range of $U^*$. Within this range, the transverse oscillations demonstrate lock-in behavior and synchronization with the vortex shedding frequency. The vorticity dynamics in the subcritical $Re$ regime reveal that in the tandem configuration, the presence of the upstream cylinder significantly modifies the wake structure, delaying vortex formation and extending the near wake. In the post-critical $Re$ regime, the cantilever shows a broader range of sustained oscillations in terms of $U^*$, with single- and multi-frequency dynamics driven by vortex-body interactions. The power transfer analysis shows cyclic energy exchange patterns between the fluid and flexible structure, with significant variations in the hydrodynamic loading along the cantilever. The findings of this work help broaden the understanding of sustained oscillations in flexible cantilevers and are relevant to the design of cantilever flow sensors. △ Less

Submitted 5 August, 2024; v1 submitted 21 September, 2023; originally announced September 2023.

A parametric level set method with convolutional encoder-decoder network for shape optimization with fluid flow

Authors: Wrik Mallik, Rajeev K. Jaiman, Jasmin Jelovica

Abstract: In this article, we present a new data-driven shape optimization approach for implicit hydrofoil morphing via a polynomial perturbation of parametric level set representation. Without introducing any change in topology, the hydrofoil morphing is achieved by six shape design variables associated with the amplitude and shape of the perturbed displacements. The proposed approach has three to four tim… ▽ More In this article, we present a new data-driven shape optimization approach for implicit hydrofoil morphing via a polynomial perturbation of parametric level set representation. Without introducing any change in topology, the hydrofoil morphing is achieved by six shape design variables associated with the amplitude and shape of the perturbed displacements. The proposed approach has three to four times lower design variables than shape optimization via free-form deformation techniques and almost two orders lower design variables compared to topology optimization via traditional parametric level sets. Using the fixed Cartesian level set mesh, we also integrate deep convolutional encoder-decoder networks as a surrogate of high-fidelity Reynolds-averaged Navier-Stokes (RANS) simulations for learning the flow field around hydrofoil shapes. We show that an efficient shape representation via parametric level sets can enable online convolutional encoder-decoder application for the shape optimization of hydrofoils. The generalized flow field prediction of the convolutional encoder-decoder is demonstrated by a mean and minimum structural similarity index measure of 0.985 and 0.95, respectively, for predicted solutions compared to RANS predictions for out-of-training shapes. The convolutional encoder-decoder predictions are performed nearly five orders of magnitude faster compared to RANS. This enables a computationally tractable surrogate-based drag minimization of fifty different hydrofoils for two different design lift coefficients. Furthermore, the best local minimum obtained via the surrogate-based optimization lie in the neighbourhood of the RANS-predicted counterparts for both the design lift cases. The present findings show promise for the future shape optimization via parametric level sets with convolutional encoder-decoder over a broader spectrum of flow conditions and shapes. △ Less

Submitted 13 January, 2023; originally announced January 2023.

A Finite Element-Inspired Hypergraph Neural Network: Application to Fluid Dynamics Simulations

Authors: Rui Gao, Indu Kant Deo, Rajeev K. Jaiman

Abstract: An emerging trend in deep learning research focuses on the applications of graph neural networks (GNNs) for mesh-based continuum mechanics simulations. Most of these learning frameworks operate on graphs wherein each edge connects two nodes. Inspired by the data connectivity in the finite element method, we present a method to construct a hypergraph by connecting the nodes by elements rather than… ▽ More An emerging trend in deep learning research focuses on the applications of graph neural networks (GNNs) for mesh-based continuum mechanics simulations. Most of these learning frameworks operate on graphs wherein each edge connects two nodes. Inspired by the data connectivity in the finite element method, we present a method to construct a hypergraph by connecting the nodes by elements rather than edges. A hypergraph message-passing network is defined on such a node-element hypergraph that mimics the calculation process of local stiffness matrices. We term this method a finite element-inspired hypergraph neural network, in short FEIH($φ$)-GNN. We further equip the proposed network with rotation equivariance, and explore its capability for modeling unsteady fluid flow systems. The effectiveness of the network is demonstrated on two common benchmark problems, namely the fluid flow around a circular cylinder and airfoil configurations. Stabilized and accurate temporal roll-out predictions can be obtained using the $φ$-GNN framework within the interpolation Reynolds number range. The network is also able to extrapolate moderately towards higher Reynolds number domain out of the training range. △ Less

Submitted 24 April, 2023; v1 submitted 29 December, 2022; originally announced December 2022.

Journal ref: Journal of Computational Physics, Volume 504, 1 May 2024, 112866

Unsteady aeroelastic characterization and scaling relations of flexible membrane wings

Authors: Guojun Li, Rajeev Kumar Jaiman

Abstract: We present a numerical study to characterize nonlinear unsteady aeroelastic interactions of two-dimensional flexible wings at high angles of attack. The coupled fluid-flexible wing system is solved by a body-fitted variational aeroelastic solver based on the fully-coupled Navier-Stokes and nonlinear structural equations. Using the coupled fluid-structure analysis, this study is aimed to provide ph… ▽ More We present a numerical study to characterize nonlinear unsteady aeroelastic interactions of two-dimensional flexible wings at high angles of attack. The coupled fluid-flexible wing system is solved by a body-fitted variational aeroelastic solver based on the fully-coupled Navier-Stokes and nonlinear structural equations. Using the coupled fluid-structure analysis, this study is aimed to provide physical insight and correlations for the aeroelastic behavior of flexible wings in the parameter space of the angle of attack and the aeroelastic number. The phase diagrams of the aerodynamic performance are established to obtain the envelope curves of the optimal performance and determine the transition line of the drag variation. The effects of the angle of attack and the aeroelastic number on the aeroelastic behaviors are systematically examined. The time-averaged membrane deformation is positively correlated with a non-dimensional number, the so-called Weber number. A new scaling relation is proposed based on the dynamic equilibrium between the aerodynamic force fluctuation and the combined inertia-elastic fluctuation. The unsteady aerodynamic force can be adjusted by manipulating the membrane vibration, the mass ratio, the Strouhal number and the aeroelastic number. The numerical investigations provide design guidelines and have the potential to enhance the maneuverability and flight agility of micro air vehicles with flexible wing structures. △ Less

Submitted 5 July, 2023; v1 submitted 22 December, 2022; originally announced December 2022.

Journal ref: AIAA Journal 2023

Self-sustained oscillations in whiskers without vortex shedding

Authors: Shayan Heydari, Mitra J. Z. Hartmann, Neelesh A. Patankar, Rajeev K. Jaiman

Abstract: Sensing the flow of water or air disturbance is critical for the survival of many animals: flow information helps them localize food, mates, and prey and to escape predators. Across species, many flow sensors take the form of long, flexible cantilevers. These cantilevers are known to exhibit sustained oscillations when interacting with fluid flow. In the presence of vortex shedding, the oscillatio… ▽ More Sensing the flow of water or air disturbance is critical for the survival of many animals: flow information helps them localize food, mates, and prey and to escape predators. Across species, many flow sensors take the form of long, flexible cantilevers. These cantilevers are known to exhibit sustained oscillations when interacting with fluid flow. In the presence of vortex shedding, the oscillations occur through mechanisms such as wake- or vortex-induced vibrations. There is, however, no clear explanation for the mechanisms governing the sustained oscillation of flexible cantilevers without vortex shedding. In recent work, we showed that a flexible cylindrical cantilever could experience sustained oscillations in its first natural vibration mode in water at Reynolds numbers below the critical Reynolds number of vortex shedding. The oscillations were shown to be driven by a frequency match (synchronization) between the flow frequency and the cantilever's first-mode natural frequency. Here, we use a body-fitted fluid-structure solver based on the Navier-Stokes and nonlinear structural equations to simulate the dynamics of a cantilevered whisker in the air at a subcritical value of Reynolds number. Results show that second-mode synchronization governs the whisker's sustained oscillation. Wavy patterns in the shear layer dominate the whisker's wake during the vibrations, indicating that parallel shear layers synchronize with the whisker's motion. As a result of this synchronization, oval-shaped motion trajectories, with matching streamwise and cross-flow vibration frequencies, are observed along the whisker. The outcomes of this study suggest possible directions for designing artificial bio-inspired flow sensors. △ Less

Submitted 8 November, 2022; originally announced November 2022.

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

Authors: Indu Kant Deo, Rui Gao, Rajeev Jaiman

Abstract: 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 phenomen… ▽ More 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 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 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. △ Less

Submitted 1 November, 2022; originally announced November 2022.

Comments: 14 pages

Predicting fluid-structure interaction with graph neural networks

Authors: Rui Gao, Rajeev K. Jaiman

Abstract: We present a rotation equivariant, quasi-monolithic graph neural network framework for the reduced-order modeling of fluid-structure interaction systems. With the aid of an arbitrary Lagrangian-Eulerian formulation, the system states are evolved temporally with two sub-networks. The movement of the mesh is reduced to the evolution of several coefficients via complex-valued proper orthogonal decomp… ▽ More We present a rotation equivariant, quasi-monolithic graph neural network framework for the reduced-order modeling of fluid-structure interaction systems. With the aid of an arbitrary Lagrangian-Eulerian formulation, the system states are evolved temporally with two sub-networks. The movement of the mesh is reduced to the evolution of several coefficients via complex-valued proper orthogonal decomposition, and the prediction of these coefficients over time is handled by a single multi-layer perceptron. A finite element-inspired hypergraph neural network is employed to predict the evolution of the fluid state based on the state of the whole system. The structural state is implicitly modeled by the movement of the mesh on the solid-fluid interface; hence it makes the proposed framework quasi-monolithic. The effectiveness of the proposed framework is assessed on two prototypical fluid-structure systems, namely the flow around an elastically-mounted cylinder, and the flow around a hyperelastic plate attached to a fixed cylinder. The proposed framework tracks the interface description and provides stable and accurate system state predictions during roll-out for at least 2000 time steps, and even demonstrates some capability in self-correcting erroneous predictions. The proposed framework also enables direct calculation of the lift and drag forces using the predicted fluid and mesh states, in contrast to existing convolution-based architectures. The proposed reduced-order model via graph neural network has implications for the development of physics-based digital twins concerning moving boundaries and fluid-structure interactions. △ Less

Submitted 16 October, 2023; v1 submitted 9 October, 2022; originally announced October 2022.

Journal ref: Physics of Fluids 36, 013622 (2024)

Assessment of convolutional recurrent autoencoder network for learning wave propagation

Authors: Wrik Mallik, Rajeev K. Jaiman, Jasmin Jelovica

Abstract: It is challenging to construct generalized physical models of wave propagation in nature owing to their complex physics as well as widely varying environmental parameters and dynamical scales. In this article, we present the convolutional autoencoder recurrent network (CRAN) as a data-driven model for learning wave propagation phenomena. The CRAN consists of a convolutional autoencoder for learnin… ▽ More It is challenging to construct generalized physical models of wave propagation in nature owing to their complex physics as well as widely varying environmental parameters and dynamical scales. In this article, we present the convolutional autoencoder recurrent network (CRAN) as a data-driven model for learning wave propagation phenomena. The CRAN consists of a convolutional autoencoder for learning low-dimensional system representation and a long short-term memory recurrent neural network for the system evolution in low dimension. We show that the convolutional autoencoder significantly outperforms the dimension-reduction of complex wave propagation phenomena via projection-based methods as it can directly learn subspaces resembling wave characteristics. On the other hand, the projection-based modes are restricted to the Fourier subspace. Geometric priors of the convolutional autoencoder enabling selective scale separation of complex wave dynamics further enhance its dimension-reduction capability. We also demonstrate that geometric priors such as translation equivariance and translational invariance of the convolutional autoencoder enable generalized learning of low-dimensional maps. Thus, the composite CRAN model connecting the convolutional autoencoder with a long short-term memory network specially designed for autoregressive modeling can perform generalized wave propagation prediction over the desired time horizon. Numerical experiments display 90% mean structural similarity index measure of CRAN predictions compared to true solutions for out-of-training cases, and less than 10% pointwise $L_1$ error for most cases, verifying such generalization claims. Finally, the CRAN predictions offer similar wave characteristic patterns to the target solutions indicating not only their generalization but also their kinematical consistency. △ Less

Submitted 11 October, 2022; v1 submitted 12 April, 2022; originally announced April 2022.

Unsteady cavitation dynamics and frequency lock-in of a freely vibrating hydrofoil at high Reynolds number

Authors: Suraj R. Kashyap, Rajeev K. Jaiman

Abstract: We investigate the influence of unsteady partial cavitation on the fluid-structure interaction of a freely vibrating hydrofoil section at high Reynolds numbers. We consider an elastically-mounted NACA66 hydrofoil section that is free to vibrate in the transverse flow direction. Cavitating flow dynamics coupled with the transverse vibration are studied at low angles of attack. For the numerical stu… ▽ More We investigate the influence of unsteady partial cavitation on the fluid-structure interaction of a freely vibrating hydrofoil section at high Reynolds numbers. We consider an elastically-mounted NACA66 hydrofoil section that is free to vibrate in the transverse flow direction. Cavitating flow dynamics coupled with the transverse vibration are studied at low angles of attack. For the numerical study, we employ a recently developed unified variational fluid-structure interaction framework based on homogeneous mixture-based cavitation with hybrid URANS-LES turbulence modeling. We first validate the numerical implementation against experimental data for turbulent cavitating flow. For freely oscillating hydrofoil, we observe large-amplitude vibrations during unsteady partial cavitating conditions that are absent in the non-cavitating flow. We identify a frequency lock-in phenomenon as the main source of sustained large-amplitude vibration whereby the unsteady lift forces lock into a sub-harmonic of the hydrofoil natural frequency. During cavity collapse and shedding, we find a periodic generation of clockwise vorticity, leading to unsteady lift generation. We determine the origin of this flow unsteadiness near the trailing edge of the hydrofoil via the interplay between the growing cavity and adverse pressure gradient. The flow-induced structural vibration is observed to have a consequent impact on the cavity dynamics. In the frequency lock-in regime, large coherent cavitating structures are seen over the hydrofoil suction surface undergoing a full cavity growth-detachment-collapse cycle. For the post-lock-in regime, cavity length is shorter and the attached cavity is observed to undergo high frequency spatially localized oscillations. In this regime, cavity shedding is primarily limited to the cavity trailing end and frequency of a complete cavity detachment and shedding event is reduced. △ Less

Submitted 22 February, 2022; originally announced February 2022.

An interface and geometry preserving phase-field method for fully Eulerian fluid-structure interaction

Authors: Xiaoyu Mao, Rajeev Jaiman

Abstract: We present an interface and geometry preserving (IGP) method for the modeling of fully Eulerian fluid-structure interaction via phase-field formulation. While the hyperbolic tangent interface profile is preserved by the time-dependent mobility model, the proposed method maintains the geometry of the solid-fluid interface by reducing the volume-conserved mean curvature flow. To achieve the reductio… ▽ More We present an interface and geometry preserving (IGP) method for the modeling of fully Eulerian fluid-structure interaction via phase-field formulation. While the hyperbolic tangent interface profile is preserved by the time-dependent mobility model, the proposed method maintains the geometry of the solid-fluid interface by reducing the volume-conserved mean curvature flow. To achieve the reduction in the curvature flow, we construct a gradient-minimizing velocity field (GMV) for the convection of the order parameter. The constructed velocity field enables the preservation of the solid velocity in the solid domain while extending the velocity in the normal direction throughout the diffuse interface region. With this treatment, the GMV reduces the normal velocity difference of the level sets of the order parameter which alleviates the undesired thickening or thinning of the diffuse interface region due to the convection. During this process, the time-dependent mobility coefficient is substantially reduced and there is a lesser curvature flow. The GMV ensures that the diffuse interface region moves with the solid bulk such that the fluid-solid interface conforms to the geometry of the solid. Using the unified momentum equation and the phase-dependent interpolation, we integrate the IGP method into a fully Eulerian variational FSI solver based on the incompressible viscous fluid and the neo-Hookean solid. We first demonstrate the ability of the phase-field-based IGP method for the convection of circular and square interfaces with a prescribed velocity field. The variational FSI framework with the IGP method is then examined for the flow passing a fixed deformable block in a channel domain. Finally, the vibration of a plate attached behind a stationary cylinder subjected to incoming flow is employed to assess the fully Eulerian framework for a large aspect ratio and sharp corners. △ Less

Submitted 27 January, 2022; originally announced January 2022.

Predicting waves in fluids with deep neural network

Authors: Indu Kant Deo, Rajeev Jaiman

Abstract: In this paper, we present a deep learning technique for data-driven predictions of wave propagation in a fluid medium. The technique relies on an attention-based convolutional recurrent autoencoder network (AB-CRAN). To construct a low-dimensional representation of wave propagation data, we employ a denoising-based convolutional autoencoder. The AB-CRAN architecture with attention-based long short… ▽ More In this paper, we present a deep learning technique for data-driven predictions of wave propagation in a fluid medium. The technique relies on an attention-based convolutional recurrent autoencoder network (AB-CRAN). To construct a low-dimensional representation of wave propagation data, we employ a denoising-based convolutional autoencoder. The AB-CRAN architecture with attention-based long short-term memory cells forms our deep neural network model for the time marching of the low-dimensional features. We assess the proposed AB-CRAN framework against the standard recurrent neural network for the low-dimensional learning of wave propagation. To demonstrate the effectiveness of the AB-CRAN model, we consider three benchmark problems, namely, one-dimensional linear convection, the nonlinear viscous Burgers equation, and the two-dimensional Saint-Venant shallow water system. Using the spatial-temporal 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 compared to the standard recurrent neural network with long short-term memory cells. The denoising autoencoder further reduces the mean squared error of prediction and improves the generalization capability in the parameter space. △ Less

Submitted 13 June, 2022; v1 submitted 17 January, 2022; originally announced January 2022.

Comments: 24 pages

Journal ref: Physics of Fluids 34, 067108 (2022)

An Interface-Driven Adaptive Variational Procedure for Fully Eulerian Fluid-Structure Interaction via Phase-field Modeling

Authors: Biswajeet Rath, Xiaoyu Mao, Rajeev K. Jaiman

Abstract: In this paper, we present a novel interface-driven adaptive variational procedure using a fully Eulerian description of fluid-structure interaction. The proposed fully-Eulerian procedure involves a fixed background unstructured mesh on which the fluid-structure interface is treated implicitly. We model the fluid-structure interaction by the phase-field finite element formulation relying on a parti… ▽ More In this paper, we present a novel interface-driven adaptive variational procedure using a fully Eulerian description of fluid-structure interaction. The proposed fully-Eulerian procedure involves a fixed background unstructured mesh on which the fluid-structure interface is treated implicitly. We model the fluid-structure interaction by the phase-field finite element formulation relying on a partitioned staggered integration of the convective Allen-Cahn equation with the unified momentum equation for both solid and fluid dynamics. We employ the positivity preserving variational scheme for a bounded and stable solution of the Allen-Cahn phase-field equation. To evaluate the solid stresses, the left Cauchy-Green deformation tensor is convected at each time step to trace the evolution of the solid strain in the Eulerian reference frame. We utilize the residual based error indicators and the newest vertex bisection algorithm for the adaptive refinement/coarsening of the unstructured mesh. The proposed nonlinear adaptive partitioned procedure restricts the coarsening step to the last non-linear iteration while simultaneously ensuring convergence properties of the coupled governing equations. We perform a detailed convergence and accuracy analysis via two benchmark problems namely, the pure solid system and a coupled fluid-solid system with an interface in a rectangle domain. We next systematically assess the performance of the adaptive procedure in terms of conservation properties for the increasing complexity of problems. Finally, we demonstrate our fully-Eulerian interface-driven adaptive FSI model to simulate the contact and bouncing phenomenon between an elastic solid and a rigid wall. △ Less

Submitted 6 February, 2022; v1 submitted 31 December, 2021; originally announced December 2021.

Deep Learning for Stability Analysis of a Freely Vibrating Sphere at Moderate Reynolds Number

Authors: A. Chizfahm, R. Jaiman

Abstract: In this paper, we present a deep learning-based reduced-order model (DL-ROM) for the stability prediction of unsteady 3D fluid-structure interaction systems. The proposed DL-ROM has the format of a nonlinear state-space model and employs a recurrent neural network with long short-term memory (LSTM). We consider a canonical fluid-structure system of an elastically-mounted sphere coupled with incomp… ▽ More In this paper, we present a deep learning-based reduced-order model (DL-ROM) for the stability prediction of unsteady 3D fluid-structure interaction systems. The proposed DL-ROM has the format of a nonlinear state-space model and employs a recurrent neural network with long short-term memory (LSTM). We consider a canonical fluid-structure system of an elastically-mounted sphere coupled with incompressible fluid flow in a state-space format. We develop a nonlinear data-driven coupling for predicting unsteady forces and vortex-induced vibration (VIV) lock-in of the freely vibrating sphere in a transverse direction. We design an input-output relationship as a temporal sequence of force and displacement datasets for a low-dimensional approximation of the fluid-structure system. Based on the prior knowledge of the VIV lock-in process, the input function contains a range of frequencies and amplitudes, which enables an efficient DL-ROM without the need for a massive training dataset for the low-dimensional modeling. Once trained, the network provides a nonlinear mapping of input-output dynamics that can predict the coupled fluid-structure dynamics for a longer horizon via the feedback process. By integrating the LSTM network with the eigensystem realization algorithm (ERA), we construct a data-driven state-space model for the reduced-order stability analysis. We investigate the underlying mechanism and stability characteristics of VIV via an eigenvalue selection process. To understand the frequency lock-in mechanism, we study the eigenvalue trajectories for a range of the reduced oscillation frequencies and the mass ratios. Consistent with the full-order simulations, the frequency lock-in branches are accurately captured by the combined LSTM-ERA procedure. The proposed DL-ROM aligns with the development of physics-based digital twin of engineering systems involving fluid-structure interactions. △ Less

Submitted 18 December, 2021; originally announced December 2021.

Comments: 33 pages, 14 figures

Three-dimensional deep learning-based reduced order model for unsteady flow dynamics with variable Reynolds number

Authors: Rachit Gupta, Rajeev Jaiman

Abstract: We present a deep learning-based reduced order model (DL-ROM) for predicting the fluid forces and unsteady vortex patterns. We consider flow past a sphere to examine the accuracy of our DL-ROM predictions. The proposed methodology relies on a three-dimensional convolutional recurrent autoencoder network (3D CRAN) to extract the low-dimensional flow features from full-order snapshots. The low-dimen… ▽ More We present a deep learning-based reduced order model (DL-ROM) for predicting the fluid forces and unsteady vortex patterns. We consider flow past a sphere to examine the accuracy of our DL-ROM predictions. The proposed methodology relies on a three-dimensional convolutional recurrent autoencoder network (3D CRAN) to extract the low-dimensional flow features from full-order snapshots. The low-dimensional features are evolved in time using a long short-term memory-based recurrent neural network and reconstructed back to the full-order as flow voxels. These 3D voxels are introduced as static and uniform query probes in the point cloud domain to reduce the unstructured mesh complexity while providing convenience in 3D CRAN training. We analyze a novel procedure to recover the interface description and the force quantities from the 3D flow voxels. The 3D CRAN methodology is first applied to an external flow past a static sphere at a single Reynolds number of Re = 300. We provide an assessment of the computing requirements in terms of the memory usage, training costs, and testing times associated with the 3D CRAN framework. Subsequently, variable Re-based flow information is infused in one 3D CRAN to learn a complicated symmetry-breaking flow regime (280 < Re < 460) for the flow past a sphere. Effects of transfer learning are analyzed for training this complicated 3D flow regime on a relatively smaller time series dataset. The 3D CRAN framework learns the flow regime nearly 20 times faster than the parallel full-order model and predicts unsteady flows with an excellent to good accuracy. Based on the predicted flow fields, the network demonstrates an R2 accuracy of 98.58% for drag and 76.43% for lift over the sphere in the chosen flow regime. The proposed framework aligns with the development of a digital twin for 3D unsteady flow field with variable Re effects. △ Less

Submitted 16 December, 2021; originally announced December 2021.

Kinematically consistent recurrent neural networks for learning inverse problems in wave propagation

Authors: Wrik Mallik, Rajeev K. Jaiman, Jasmin Jelovica

Abstract: Although machine learning (ML) is increasingly employed recently for mechanistic problems, the black-box nature of conventional ML architectures lacks the physical knowledge to infer unforeseen input conditions. This implies both severe overfitting during a dearth of training data and inadequate physical interpretability, which motivates us to propose a new kinematically consistent, physics-based… ▽ More Although machine learning (ML) is increasingly employed recently for mechanistic problems, the black-box nature of conventional ML architectures lacks the physical knowledge to infer unforeseen input conditions. This implies both severe overfitting during a dearth of training data and inadequate physical interpretability, which motivates us to propose a new kinematically consistent, physics-based ML model. In particular, we attempt to perform physically interpretable learning of inverse problems in wave propagation without suffering overfitting restrictions. Towards this goal, we employ long short-term memory (LSTM) networks endowed with a physical, hyperparameter-driven regularizer, performing penalty-based enforcement of the characteristic geometries. Since these characteristics are the kinematical invariances of wave propagation phenomena, maintaining their structure provides kinematical consistency to the network. Even with modest training data, the kinematically consistent network can reduce the $L_1$ and $L_\infty$ error norms of the plain LSTM predictions by about 45% and 55%, respectively. It can also increase the horizon of the plain LSTM's forecasting by almost two times. To achieve this, an optimal range of the physical hyperparameter, analogous to an artificial bulk modulus, has been established through numerical experiments. The efficacy of the proposed method in alleviating overfitting, and the physical interpretability of the learning mechanism, are also discussed. Such an application of kinematically consistent LSTM networks for wave propagation learning is presented here for the first time. △ Less

Submitted 8 October, 2021; originally announced October 2021.

On the fluid-structure interaction of a flexible cantilever cylinder at low Reynolds numbers

Authors: S. Heydari, N. A. Patankar, M. J. Z. Hartmann, R. K. Jaiman

Abstract: We present a numerical study to investigate the fluid-structure interaction of a flexible circular cantilever cylinder in a uniform cross-flow. We employ a fully-coupled fluid-structure solver based on the three-dimensional Navier-Stokes equations and the Euler-Bernoulli beam theory. We examine the dynamics of the cylinder for a wide range of reduced velocities ($U^*$), mass ratios ($m^*$), and Re… ▽ More We present a numerical study to investigate the fluid-structure interaction of a flexible circular cantilever cylinder in a uniform cross-flow. We employ a fully-coupled fluid-structure solver based on the three-dimensional Navier-Stokes equations and the Euler-Bernoulli beam theory. We examine the dynamics of the cylinder for a wide range of reduced velocities ($U^*$), mass ratios ($m^*$), and Reynolds numbers ($Re$). Of particular interest is to explore the possibility of flow-induced vibrations in a slender cantilever cylinder of aspect ratio $AR=100$ at laminar subcritical $Re$ regime (i.e., no periodic vortex shedding). We assess the extent to which such a flexible cylindrical beam can sustain flow-induced vibrations and characterize the contribution of the beam's flexibility to the stability of the wake at low $Re$. We show that when certain conditions are satisfied, the flexible cantilever cylinder undergoes sustained large-amplitude vibrations. The frequency of the oscillations is found to match the frequency of the periodic fluid forces for a particular range of system parameters. In this range, the frequency of the transverse vibrations is shown to match the first-mode natural frequency of the cylinder, indicating the existence of the lock-in phenomenon. The range of the lock-in regime is shown to have a strong dependence on $Re$ and $m^*$. We discover that unlike the steady wake behind a stationary rigid cylinder, the wake of a low mass ratio flexible cantilever cylinder could lose its stability in the lock-in regime at Reynolds numbers as low as $Re=22$. A combined VIV-galloping type instability is shown to be the possible cause of the wake instability at this $Re$ regime. These findings attempt to generalize our understanding of the flow-induced vibrations in flexible cantilever structures and can have a profound impact on the development of novel flow-measurement sensors. △ Less

Submitted 25 May, 2021; originally announced May 2021.

A high-fidelity numerical study on the propulsive performance of pitching flexible plates

Authors: Guojun Li, Gael Kemp, Rajeev Kumar Jaiman, Boo Cheong Khoo

Abstract: In this paper, we numerically investigate the propulsive performance of three-dimensional pitching flexible plates with varying flexibility and trailing edge shapes. To eliminate the effect of other geometric parameters, only the trailing edge angle is varied from 45° (concave), 90° (rectangular) to 135° (convex) while maintaining the constant area of the flexible plate. We examine the impact of t… ▽ More In this paper, we numerically investigate the propulsive performance of three-dimensional pitching flexible plates with varying flexibility and trailing edge shapes. To eliminate the effect of other geometric parameters, only the trailing edge angle is varied from 45° (concave), 90° (rectangular) to 135° (convex) while maintaining the constant area of the flexible plate. We examine the impact of the frequency ratio f* defined as the ratio of the natural frequency of the flexible plate to the actuated pitching frequency. Through our numerical simulations, we find that the global maximum mean thrust occurs near f*=1 corresponding to the resonance condition. However, the optimal propulsive efficiency is achieved around f*=1.54 instead of the resonance condition. While the convex plate with low and high bending stiffness values shows the best performance, the rectangular plate with moderate bending stiffness is the most efficient propulsion configuration. Through dynamic mode decomposition, we find that the passive deformation can help in redistributing the pressure gradient thus improving the efficiency and thrust production. A momentum-based thrust evaluation approach is adopted to link the instantaneous vortical structures with the time-dependent thrust. When the vortices detach from the trailing edge, the instantaneous thrust shows the largest values due to the strong momentum change and convection process. Moderate flexibility and convex shape help transfer momentum to the fluid, thereby improving thrust generation and promoting the transition from drag to thrust. The increase of the trailing edge angle can broaden the range of flexibility that produces positive mean thrust. △ Less

Submitted 7 October, 2021; v1 submitted 2 March, 2021; originally announced March 2021.

A robust and accurate finite element framework for cavitating flows with fluid-structure interaction

Authors: Suraj R. Kashyap, Rajeev K. Jaiman

Abstract: We present a unified variational mechanics framework for cavitating turbulent flows and structural motions via a stabilized finite element formulation. To model the finite mass transfer rate in cavitation phenomena, we employ the homogenous mixture-based approach via phenomenological scalar transport differential equations. Stable linearizations of the finite mass transfer terms for the mass conti… ▽ More We present a unified variational mechanics framework for cavitating turbulent flows and structural motions via a stabilized finite element formulation. To model the finite mass transfer rate in cavitation phenomena, we employ the homogenous mixture-based approach via phenomenological scalar transport differential equations. Stable linearizations of the finite mass transfer terms for the mass continuity equation and the scalar transport equations are derived for robust and accurate implementation. The linearized matrices for the cavitation equation are imparted a positivity-preserving property to address numerical oscillations arising from high-density gradients typical of two-phase cavitating flows. The proposed formulation is strongly coupled in a partitioned manner with an incompressible 3D Navier-Stokes finite element solver, and the unsteady problem is advanced in time using a fully-implicit generalized-$α$ time integration scheme. We first verify the implementation on the benchmark case of Rayleigh bubble collapse. We demonstrate the accuracy and convergence of the cavitation solver by comparing the numerical solutions with the analytical solutions of the Rayleigh-Plesset equation for bubble dynamics. We find our solver to be robust for large time steps and the absence of spurious oscillations in the pressure field. The cavitating flow solver is coupled with a hybrid URANS-LES turbulence model with a turbulence viscosity corrected for the presence of vapor. We validate the coupled solver for a very high Reynolds number turbulent cavitating flow over a NACA0012 hydrofoil section. Finally, the proposed method is solved in an Arbitrary Lagrangian-Eulerian framework to study turbulent cavitating flow over a pitching hydrofoil section and the coupled FSI results are explored for characteristic features of cavitating flows such as re-entrant jet and periodic cavity shedding. △ Less

Submitted 18 February, 2021; originally announced February 2021.

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

Authors: Rachit Gupta, Rajeev Kumar Jaiman

Abstract: We present a hybrid partitioned deep learning framework for the reduced-order modeling of fluid-structure interaction. Using the discretized Navier-Stokes in the arbitrary Lagrangian-Eulerian reference frame, we generate the full-order flow snapshots and point cloud displacements as target data for the learning and inference of fluid-structure dynamics. The hybrid operation of this methodology com… ▽ More We present a hybrid partitioned deep learning framework for the reduced-order modeling of fluid-structure interaction. Using the discretized Navier-Stokes in the arbitrary Lagrangian-Eulerian reference frame, we generate the full-order flow snapshots and point cloud displacements as target data for the learning and inference of fluid-structure dynamics. The hybrid operation of this methodology comes by combining two data-driven models for fluid and solid subdomains via deep learning-based reduced-order models (DL-ROMs). The proposed framework comprises the partitioned data-driven drivers for unsteady flow and the moving point cloud displacements. At the fluid-structure interface, the force information is exchanged between the two partitioned subdomain solvers. The first component of our framework relies on the proper orthogonal decomposition-based recurrent neural network (POD-RNN) as a DL-ROM procedure to infer the point cloud with a moving interface. This model utilizes the POD basis modes to reduce dimensionality and evolve them in time via RNN. The second component employs the convolution-based recurrent autoencoder network (CRAN) as a self-supervised DL-ROM procedure to infer the nonlinear flow dynamics at static Eulerian probes. We introduce these probes as spatially structured query nodes in the moving point cloud to treat the Lagrangian-to-Eulerian conflict together with convenience in training the CRAN driver. To determine these Eulerian probes, we construct a novel snapshot-field transfer and load recovery algorithm. A popular prototypical fluid-structure interaction problem of flow past a freely oscillating cylinder is considered to assess the efficacy of the proposed methodology for a different set of reduced velocities. The proposed framework tracks the interface description with acceptable accuracy and predicts the nonlinear wake dynamics over the chosen test data range. △ Less

Submitted 30 October, 2021; v1 submitted 17 February, 2021; originally announced February 2021.

Flow-excited membrane instability at moderate Reynolds numbers

Authors: Guojun Li, Rajeev Kumar Jaiman, Boo Cheong Khoo

Abstract: In this paper, we study the fluid-structure interaction (FSI) of a three-dimensional (3D) flexible membrane immersed in an unsteady separated flow at moderate Reynolds numbers. We employ a body-conforming variational FSI solver based on the recently developed partitioned iterative scheme for the coupling of turbulent fluid flow with nonlinear structural dynamics. Of particular interest is to under… ▽ More In this paper, we study the fluid-structure interaction (FSI) of a three-dimensional (3D) flexible membrane immersed in an unsteady separated flow at moderate Reynolds numbers. We employ a body-conforming variational FSI solver based on the recently developed partitioned iterative scheme for the coupling of turbulent fluid flow with nonlinear structural dynamics. Of particular interest is to understand the flow-excited instability of a 3D flexible membrane as a function of the non-dimensional mass ratio, Reynolds number and aeroelastic number. For a wide range of the parameters, we examine two distinctive stability regimes of fluid-membrane interaction: deformed-steady state (DSS) and dynamic balance state (DBS). We propose stability phase diagrams to demarcate the DSS and DBS regimes for the parameter space of mass ratio vs. Reynolds number and mass ratio vs. aeroelastic number. Based on the aeroelastic mode analysis, we observe a frequency synchronization between the vortex shedding frequency and the membrane vibration frequency which leads to self-sustained vibrations in the dynamic balance state. To characterize the origin of the frequency lock-in, we derive an approximate analytical formula for the nonlinear natural frequency by considering the added mass effect and employing a large deflection theory for a simply supported rectangular membrane. Through our systematic high-fidelity numerical investigation, we find that the onset of the membrane vibration and the mode transition has a dependence on the frequency lock-in between the natural frequency of the tensioned membrane and the vortex shedding frequency or its harmonics. These findings on the fluid-elastic instability of membranes have implications for the design and development of control strategies for membrane wing-based unmanned systems and drones. △ Less

Submitted 7 October, 2021; v1 submitted 20 November, 2020; originally announced November 2020.

Comments: arXiv admin note: text overlap with arXiv:2006.13467

Journal ref: J. Fluid Mech. 929 (2021) A40

Assessment of unsteady flow predictions using hybrid deep learning based reduced order models

Authors: Sandeep Reddy Bukka, Rachit Gupta, Allan Ross Magee, Rajeev Kumar Jaiman

Abstract: In this paper, we present two deep learning-based hybrid data-driven reduced order models for the prediction of unsteady fluid flows. The first model projects the high-fidelity time series data from a finite element Navier-Stokes solver to a low-dimensional subspace via proper orthogonal decomposition (POD). The time-dependent coefficients in the POD subspace are propagated by the recurrent net (c… ▽ More In this paper, we present two deep learning-based hybrid data-driven reduced order models for the prediction of unsteady fluid flows. The first model projects the high-fidelity time series data from a finite element Navier-Stokes solver to a low-dimensional subspace via proper orthogonal decomposition (POD). The time-dependent coefficients in the POD subspace are propagated by the recurrent net (closed-loop encoder-decoder updates) and mapped to a high-dimensional state via the mean flow field and POD basis vectors. This model is referred as POD-RNN. The second model, referred to as convolution recurrent autoencoder network (CRAN), employs convolutional neural networks (CNN) as layers of linear kernels with nonlinear activations, to extract low-dimensional features from flow field snapshots. The flattened features are advanced using a recurrent (closed-loop manner) net and up-sampled (transpose convoluted) gradually to high-dimensional snapshots. Two benchmark problems of the flow past a cylinder and flow past a side-by-side cylinder are selected as the test problems to assess the efficacy of these models. For the problem of flow past a single cylinder, the performance of both the models is satisfactory, with CRAN being a bit overkill. However, it completely outperforms the POD-RNN model for a more complicated problem of flow past side-by-side cylinders. Owing to the scalability of CRAN, we briefly introduce an observer-corrector method for the calculation of integrated pressure force coefficients on the fluid-solid boundary on a reference grid. This reference grid, typically a structured and uniform grid, is used to interpolate scattered high-dimensional field data as snapshot images. These input images are convenient in training CRAN. This motivates us to further explore the application of CRAN models for the prediction of fluid flows. △ Less

Submitted 22 September, 2020; v1 submitted 3 September, 2020; originally announced September 2020.

A variational interface-preserving and conservative phase-field method for the surface tension effect in two-phase flows

Authors: Xiaoyu Mao, Vaibhav Joshi, Rajeev Jaiman

Abstract: We present a finite element based variational interface-preserving and conservative phase-field formulation for the modeling of incompressible two-phase flows with surface tension dynamics. The preservation of the hyperbolic tangent interface profile of the convective Allen-Cahn phase-field formulation relies on a novel time-dependent mobility model. The mobility coefficient is adjusted adaptively… ▽ More We present a finite element based variational interface-preserving and conservative phase-field formulation for the modeling of incompressible two-phase flows with surface tension dynamics. The preservation of the hyperbolic tangent interface profile of the convective Allen-Cahn phase-field formulation relies on a novel time-dependent mobility model. The mobility coefficient is adjusted adaptively as a function of gradients of the velocity and the order parameter in the diffuse interface region in such a way that the free energy minimization properly opposes the convective distortion. The ratio of the convective distortion to the free energy minimization is termed as the convective distortion parameter, which characterizes the deviation from the hyperbolic tangent shape due to the convection effect. The mass conservation is achieved by enforcing a Lagrange multiplier with both temporal and spatial dependence on the phase-field function. We integrate the interface-preserving and conservative phase-field formulation with the incompressible Navier-Stokes equations and the continuum surface tension force model for the simulation of incompressible two-phase flows. A positivity preserving scheme designed for the boundedness and stability of the solution is employed for the variational discretization using unstructured finite elements. We examine the convergence and accuracy of the Allen-Cahn phase-field solver through a generic one-dimensional bistable diffusion-reaction system in a stretching flow. We quantify and systematically assess the relative interface thickness error and the relative surface tension force error with respect to the convective distortion parameter. Two- and three-dimensional rising bubble cases are further simulated to examine the effectiveness of the proposed model on the volume-preserving mean curvature flow and the interface-preserving capability. △ Less

Submitted 31 July, 2020; originally announced July 2020.

Aeroelastic mode decomposition framework and mode selection mechanism in fluid-membrane interaction

Authors: G. Li, R. K. Jaiman, B. C. Khoo

Abstract: In this study, we present a global Fourier mode decomposition framework for unsteady fluid-structure interaction. We apply the framework to isolate and extract the aeroelastic modes arising from a coupled three-dimensional fluid-membrane system. The proposed framework is employed to decompose the physical variables in the fluid and structural domains into frequency-ranked aeroelastic modes in a un… ▽ More In this study, we present a global Fourier mode decomposition framework for unsteady fluid-structure interaction. We apply the framework to isolate and extract the aeroelastic modes arising from a coupled three-dimensional fluid-membrane system. The proposed framework is employed to decompose the physical variables in the fluid and structural domains into frequency-ranked aeroelastic modes in a unified way. We observe the frequency synchronization between the vortex shedding and the structural vibration via mode decomposition analysis. We examine the role of flexibility in the aeroelastic mode selection and perform a systematic comparison of flow features among a rigid wing, a rigid cambered wing and a flexible membrane. With the aid of our mode decomposition technique, we find that the dominant structural mode exhibits a chordwise second and spanwise first mode at different angles of attack. The structural natural frequency corresponding to this mode is estimated using an approximate analytical formula. By examining the dominant frequency of the coupled system, we find that the dominant membrane vibrational mode is selected via the frequency lock-in between the dominant vortex shedding frequency and the structural natural frequency. From the fluid modes and the mode energy spectra at $α=20^\circ$ and $25^\circ$, the aeroelastic modes corresponding to the non-integer frequency components lower than the dominant frequency are found to be associated with the bluff body vortex shedding instability. The non-periodic aeroelastic response observed at higher angles of attack are related to the interaction between aeroelastic modes caused by the frequency lock-in and the bluff-body-like vortex shedding. △ Less

Submitted 12 July, 2021; v1 submitted 24 June, 2020; originally announced June 2020.

Transverse flow-induced vibrations of a sphere in the proximity of a free surface: A numerical study

Authors: Amir Chizfahm, Vaibhav Joshi, Rajeev Jaiman

Abstract: We present a numerical study on the transverse flow-induced vibration (FIV) of an elastically mounted sphere in the vicinity of a free surface at subcritical Reynolds numbers. To begin, We verify and analyze the mode transitions and the motion trajectories of a fully submerged sphere vibrating freely in all directions for the Reynolds number up to $30\,000$. Next, the response dynamics of a transv… ▽ More We present a numerical study on the transverse flow-induced vibration (FIV) of an elastically mounted sphere in the vicinity of a free surface at subcritical Reynolds numbers. To begin, We verify and analyze the mode transitions and the motion trajectories of a fully submerged sphere vibrating freely in all directions for the Reynolds number up to $30\,000$. Next, the response dynamics of a transversely vibrating sphere is studied for three values of normalized immersion ratio ($h^*=h/D$, where $h$ is the distance from the top of the sphere to undisturbed free-surface level and $D$ is the sphere diameter), at $h^*=1$ (fully submerged sphere with no free-surface effect), $h^*=0$ (where the top of the sphere touches the free surface) and $h^*=-0.25$ (where the sphere pierces the free surface). At the lock-in range, we observe that the amplitude response at $h^*=0$ is decreased significantly compared to the case at $h^*=1$. It is found that the vorticity flux is diffused due to the free-surface boundary and the free surface acts as a sink of energy that leads to a reduction in the transverse force and amplitude response. When the sphere pierces the free surface at $h^*=-0.25$, the amplitude response at the lock-in state is found to be greater than all the submerged cases studied with the maximum peak-to-peak amplitude of $\sim2D$. We find that the interaction of the piercing sphere with the air-water interface causes a relatively large surface deformation and has a significant impact on the synchronization of the vortex shedding and the vibration frequency. Increased streamwise vorticity gives rise to a relatively larger transverse force to the piercing sphere at $h^*=-0.25$, resulting in greater positive energy transfer per cycle to sustain the large-amplitude vibration. Lasty, we study the sensitivity of FIV response on the mass ratio, $m^*$, and Froude number, $Fr$, at the lock-in state. △ Less

Submitted 26 May, 2020; originally announced May 2020.

Comments: 45 pages, 32 figures

A variational flexible multibody formulation for partitioned fluid-structure interaction: Application to bat-inspired drones and unmanned air-vehicles

Authors: Vaibhav Joshi, Rajeev K. Jaiman, Carl Ollivier-Gooch

Abstract: We present a three-dimensional (3D) partitioned aeroelastic formulation for a flexible multibody system interacting with incompressible turbulent fluid flow. While the incompressible Navier-Stokes system is discretized using a stabilized Petrov-Galerkin procedure, the multibody structural system consists of a generic interaction of multiple components such as rigid body, beams and flexible thin sh… ▽ More We present a three-dimensional (3D) partitioned aeroelastic formulation for a flexible multibody system interacting with incompressible turbulent fluid flow. While the incompressible Navier-Stokes system is discretized using a stabilized Petrov-Galerkin procedure, the multibody structural system consists of a generic interaction of multiple components such as rigid body, beams and flexible thin shells along with various types of joints and connections among them. A co-rotational framework is utilized for the category of small strain problems where the displacement of the body is decomposed into a rigid body rotation and a small strain component. This assumption simplifies the structural equations and allows for the incorporation of multiple bodies (rigid as well as flexible) in the system. The displacement and rotation constraints at the joints are imposed by a Lagrange multiplier method. The equilibrium conditions at the fluid-structure interface are satisfied by the transfer of tractions and structural displacements via the radial basis function approach, which is globally conservative. For the coupled stability in low structure-to-fluid mass ratio regimes, a nonlinear iterative force correction scheme is employed in the partitioned staggered predictor-corrector scheme. The convergence and generality of the radial basis function mapping are analyzed by carrying out systematic error analysis of the transfer of fluid traction across the non-matching fluid-structure interface where a third-order of convergence is observed. The proposed aeroelastic framework is then validated by considering a flow across a flexible pitching plate configuration with serration at the trailing edge. Finally, we demonstrate the flow across a flexible flapping wing of a bat modeling the bone fingers as beams and the flexible membrane as thin shells in the multibody system along with the joints. △ Less

Submitted 5 May, 2020; originally announced May 2020.

Comments: 34 pages, 20 figures

Deep Convolutional Recurrent Autoencoders for Flow Field Prediction

Authors: Sandeep Reddy Bukka, Allan Ross Magee, Rajeev Kumar Jaiman

Abstract: In this paper, an end-to-end nonlinear model reduction methodology is presented based on the convolutional recurrent autoencoder networks. The methodology is developed in the context of the overall data-driven reduced-order model framework proposed in the paper. The basic idea behind the methodology is to obtain the low dimensional representations via convolutional neural networks and evolve these… ▽ More In this paper, an end-to-end nonlinear model reduction methodology is presented based on the convolutional recurrent autoencoder networks. The methodology is developed in the context of the overall data-driven reduced-order model framework proposed in the paper. The basic idea behind the methodology is to obtain the low dimensional representations via convolutional neural networks and evolve these low dimensional features via recurrent neural networks in the time domain. The high dimensional representations are constructed from the evolved low dimensional features via transpose convolutional neural networks. With an unsupervised training strategy, the model serves as an end to end tool which can evolve the flow state of the nonlinear dynamical system. The convolutional recurrent autoencoder network model is applied to the problem of flow past bluff bodies for the first time. To demonstrate the effectiveness of the methodology, two canonical problems namely the flow past a plain cylinder and the flow past side-by-side cylinders are explored in this paper. Pressure and velocity fields of the unsteady flow are predicted in future via the convolutional recurrent autoencoder model. The performance of the model is satisfactory for both the problems. Specifically, the multiscale nature and the gap flow dynamics of the side-by-side cylinders are captured by the proposed data-driven model reduction methodology. The error metrics, the normalized squared error, and the normalized reconstruction error are considered for the assessment of the data-driven framework. △ Less

Submitted 22 March, 2020; originally announced March 2020.

Comments: arXiv admin note: text overlap with arXiv:1808.01346 by other authors

A projection-based numerical integration scheme for embedded interface: Application to fluid-structure interaction

Authors: Bin Liu, Rajeev Kumar Jaiman, Danielle Sweimann Tan

Abstract: We present a projection-based numerical integration technique to deal with embedded interface in finite element (FE) framework. The element cut by an embedded interface is denoted as a cut cell. We recognize elemental matrices of a cut cell can be reconstructed from the elemental matrices of its sub-divided cells, via projection at matrix level. These sub-divided cells are termed as integration ce… ▽ More We present a projection-based numerical integration technique to deal with embedded interface in finite element (FE) framework. The element cut by an embedded interface is denoted as a cut cell. We recognize elemental matrices of a cut cell can be reconstructed from the elemental matrices of its sub-divided cells, via projection at matrix level. These sub-divided cells are termed as integration cells. The proposed technique possesses following characteristics (1) no change in FE formulation and quadrature rule; (2) consistency with the derivation of FE formulation in variational principle. It can be considered as a re-projection of the residuals of equation system in the test function space or a reduced-order modeling (ROM) technique. These characteristics significantly improves its scalability, easy-to-implementation and robustness to deal with problems involving embedded discontinuities in FE framework. Numerical examples, e.g., vortex-induced vibration (VIV), rotation, free fall and rigid-body contact in which the proposed technique is implemented to integrate the variational form of Navier-Stokes equations in cut cells, are presented. △ Less

Submitted 17 November, 2018; originally announced November 2018.

A Hybrid Data-driven Deep Learning Technique for Fluid-Structure Interaction

Authors: T. P. Miyanawala, R. K. Jaiman

Abstract: This paper is concerned with the development of a hybrid data-driven technique for unsteady fluid-structure interaction systems. The proposed data-driven technique combines the deep learning framework with a projection-based low-order modeling. While the deep learning provides low-dimensional approximations from datasets arising from black-box solvers, the projection-based model constructs the low… ▽ More This paper is concerned with the development of a hybrid data-driven technique for unsteady fluid-structure interaction systems. The proposed data-driven technique combines the deep learning framework with a projection-based low-order modeling. While the deep learning provides low-dimensional approximations from datasets arising from black-box solvers, the projection-based model constructs the low-dimensional approximations by projecting the original high-dimensional model onto a low-dimensional subspace. Of particular interest of this paper is to predict the long time series of unsteady flow fields of a freely vibrating bluff-body subjected to wake-body synchronization. We consider convolutional neural networks (CNN) for the learning dynamics of wake-body interaction, which assemble layers of linear convolutions with nonlinear activations to automatically extract the low-dimensional flow features. Using the high-fidelity time series data from the stabilized finite element Navier-Stokes solver, we first project the dataset to a low-dimensional subspace by proper orthogonal decomposition (POD) technique. The time-dependent coefficients of the POD subspace are mapped to the flow field via a CNN with nonlinear rectification, and the CNN is iteratively trained using the stochastic gradient descent method to predict the POD time coefficient when a new flow field is fed to it. The time-averaged flow field, the POD basis vectors, and the trained CNN are used to predict the long time series of the flow fields and the flow predictions are quantitatively assessed with the full-order (high-dimensional) simulation data. The proposed POD-CNN model based on the data-driven approximation has a remarkable accuracy in the entire fluid domain including the highly nonlinear near wake region behind a freely vibrating bluff body. △ Less

Submitted 13 February, 2019; v1 submitted 19 July, 2018; originally announced July 2018.

Comments: 10 pages, 6 figures, under consideration to be published in OMAE 2019

A novel 3D variational aeroelastic framework for flexible multibody dynamics: Application to bat-like flapping dynamics

Authors: Guojun Li, Yun Zhi Law, Rajeev K. Jaiman

Abstract: We present a novel three-dimensional (3D) variational aeroelastic framework for flapping wing with a flexible multibody system subjected to an external incompressible turbulent flow. The proposed aeroelastic framework consists of a three-dimensional fluid solver with delayed detached eddy simulation (DDES) and a nonlinear monolithic elastic structural solver for the flexible multibody system with… ▽ More We present a novel three-dimensional (3D) variational aeroelastic framework for flapping wing with a flexible multibody system subjected to an external incompressible turbulent flow. The proposed aeroelastic framework consists of a three-dimensional fluid solver with delayed detached eddy simulation (DDES) and a nonlinear monolithic elastic structural solver for the flexible multibody system with constraints. Radial basis function (RBF) is applied in this framework to transfer the aerodynamic forces and structural displacements across the discrete non-matching interface meshes while satisfying a global energy conservation. The fluid equations are discretized using a stabilized Petrov-Galerkin method in space and the generalized-$α$ approach is employed to integrate the solution in time. The flexible multibody system is solved by using geometrically exact co-rotational finite element method and an energy decaying scheme is used to achieve numerical stability of the multibody solver with constraints. A nonlinear iterative force correction (NIFC) scheme is applied in a staggered partitioned iterative manner to maintain the numerical stability of aeroelastic coupling with strong added mass effect. An isotropic aluminum wing with flapping motion is simulated via the proposed aeroelastic framework and the accuracy of the coupled solution is validated with the available experimental data. We next study the robustness and reliability of the 3D flexible multibody aeroelastic framework for an anisotropic flapping wing flight involving battens and membranes with composite material and compare against the experimental results. Finally, we demonstrate the aeroelastic framework for a bat-like wing and examine the effects of flexibility on the flapping wing dynamics. △ Less

Submitted 11 July, 2018; originally announced July 2018.

Comments: 53 pages

Journal ref: Computers & Fluids 2019

Decomposition of wake dynamics in fluid-structure interaction via low-dimensional models

Authors: Tharindu P. Miyanawala, Rajeev K. Jaiman

Abstract: We present a dynamic decomposition analysis of the wake flow in fluid-structure interaction (FSI) systems under both laminar and turbulent flow conditions. Of particular interest is to provide the significance of low-dimensional wake flow features and their interaction dynamics to sustain the free vibration of a square cylinder at a relatively low mass ratio. To obtain the high-dimensional data, w… ▽ More We present a dynamic decomposition analysis of the wake flow in fluid-structure interaction (FSI) systems under both laminar and turbulent flow conditions. Of particular interest is to provide the significance of low-dimensional wake flow features and their interaction dynamics to sustain the free vibration of a square cylinder at a relatively low mass ratio. To obtain the high-dimensional data, we employ a body-conforming variational fluid-structure interaction solver based on the recently developed partitioned iterative scheme and the dynamic subgrid-scale turbulence model for a moderate Reynolds number. The snapshot data from high-dimensional FSI simulations are projected to a low-dimensional subspace using the proper orthogonal decomposition (POD). We utilize each corresponding POD mode for detecting features: the vortex street, the shear layer and the near-wake bubble. We find that the vortex shedding modes contribute solely to the lift force, while the near-wake and shear layer modes play a dominating role to the drag force. We further examine the fundamental mechanism of this dynamical behavior and propose a force decomposition technique via low-dimensional approximation. We ascertain quantitatively that the shear layer feeds the vorticity flux to the wake vortices and the near-wake bubble during the wake-body synchronization. Based on the decomposition of wake dynamics, we suggest an interaction cycle for the frequency lock-in during the wake-body interaction, which provides the inter-relationship between the high amplitude motion and the dominating wake features. Through our investigation of wake-body synchronization below critical Re range, we discover that the bluff body can undergo a synchronized high-amplitude vibration due to flexibility-induced unsteadiness. The interaction cycle for the wake-synchronization is found to be valid for the turbulent wake flow. △ Less

Submitted 1 June, 2018; originally announced June 2018.

Comments: 42 pages, 20 figures

Journal ref: Journal of Fluid Mechanics. 2019 May;867:723-64

A hybrid variational Allen-Cahn/ALE scheme for the coupled analysis of two-phase fluid-structure interaction

Authors: Vaibhav Joshi, Rajeev K. Jaiman

Abstract: We present a novel partitioned iterative formulation for modeling of fluid-structure interaction in two-phase flows. The variational formulation consists of a stable and robust integration of three blocks of differential equations, viz., incompressible viscous fluid, a rigid or flexible structure and two-phase indicator field. The fluid-fluid interface between the two phases, which may have high d… ▽ More We present a novel partitioned iterative formulation for modeling of fluid-structure interaction in two-phase flows. The variational formulation consists of a stable and robust integration of three blocks of differential equations, viz., incompressible viscous fluid, a rigid or flexible structure and two-phase indicator field. The fluid-fluid interface between the two phases, which may have high density and viscosity ratios, is evolved by solving the conservative phase-field Allen-Cahn equation in the arbitrary Lagrangian-Eulerian coordinates. While the Navier-Stokes equations are solved by a stabilized Petrov-Galerkin method, the conservative Allen-Chan phase-field equation is discretized by the positivity preserving variational scheme. Fully decoupled implicit solvers for the two-phase fluid and the structure are integrated by the nonlinear iterative force correction in a staggered partitioned manner. We assess the accuracy and stability of the new phase-field/ALE variational formulation for two- and three-dimensional problems involving the dynamical interaction of rigid bodies with free-surface. We consider the decay test problems of increasing complexity, namely free translational heave decay of a circular cylinder and free rotation of a rectangular barge. Through numerical experiments, we show that the proposed formulation is stable and robust for high density ratios across fluid-fluid interface and for low structure-to-fluid mass ratio with strong added-mass effects. Using three-dimensional unstructured meshes, we demonstrate the second-order temporal accuracy of the coupled phase-field/ALE method. Finally, we demonstrate the three-dimensional phase-field FSI formulation for a practical problem of internal two-phase flow in a flexible circular pipe subjected to vortex-induced vibrations due to external fluid flow. △ Less

Submitted 21 March, 2018; originally announced March 2018.

Comments: 25 pages, 14 figures, 1 table

On the turbulent flow field over riblets of various groove sizes at low Reynolds number

Authors: J. H. Ng, R. K. Jaiman, T. T. Lim

Abstract: In this work, low Reynolds number turbulent flow through a corrugated channel, formed by standard V-groove riblets, are investigated via direct numerical simulations (DNS). The simulations attempt to assess the variation of flow dynamics caused by a change in the characteristic size of the V-groove riblets. Such characterization of size effects provides a means to identify the local flow features… ▽ More In this work, low Reynolds number turbulent flow through a corrugated channel, formed by standard V-groove riblets, are investigated via direct numerical simulations (DNS). The simulations attempt to assess the variation of flow dynamics caused by a change in the characteristic size of the V-groove riblets. Such characterization of size effects provides a means to identify the local flow features arising from fluid-riblet interaction, and to investigate their relevance to the change in viscous drag and turbulence statistics. The present simulations confirmed an improved variation trend when the flow dynamics is examined in the light of the groove size. At the lower end of the range of groove size considered, the profiles of turbulence statistics more or less resemble the plane channel flow, except for a systematic shift with the groove size. When the groove size becomes increasingly large, the lodging of near-wall flow structures and the generation of mean secondary flows tend to be more apparent. The collective impact of the lodging of flow structures and the mean secondary flow correlates with the increase in viscous drag, and leads to significant alterations of the turbulence statistics. △ Less

Submitted 20 March, 2018; originally announced March 2018.

Dynamics of gap flow interference in a vibrating side-by-side arrangement of two circular cylinders at moderate Reynolds number

Authors: B. Liu, R. K. Jaiman

Abstract: In this work, the coupled dynamics of the gap flow and the vortex-induced vibration (VIV) on a side-by-side (SBS) arrangement of two circular cylinders is numerically investigated at moderate Reynolds number 100 < Re < 800. Of particular interest is to establish a relationship between the VIV, the gap flow and the near-wake instability behind bluff bodies. We find that the kinematics of the VIV re… ▽ More In this work, the coupled dynamics of the gap flow and the vortex-induced vibration (VIV) on a side-by-side (SBS) arrangement of two circular cylinders is numerically investigated at moderate Reynolds number 100 < Re < 800. Of particular interest is to establish a relationship between the VIV, the gap flow and the near-wake instability behind bluff bodies. We find that the kinematics of the VIV regulates the streamwise vorticity concentration, which accompanies with a recovery of two-dimensional hydrodynamic responses at the peak lock-in stage. On the other hand, the near-wake instability may develop around an in-determinant two-dimensional streamline saddle point along the interfaces of a pair of imbalanced counter-signed vorticity clusters. The vorticity concentration difference of adjacent vorticity clusters and the fluid momentum are closely interlinked with the prominence of streamwise vortical structures. In both SSBS and VSBS arrangements, the flip-flopping frequency is significantly low for the three-dimensional flow, except at the VIV lock-in stage for the VSBS arrangement. A quasi-stable deflected gap flow regime with negligible spanwise hydrodynamic (i.e., two-dimensional) response is found at the peak lock-in stage of VSBS arrangements. Owing to the gap-flow proximity interference, a high streamwise vorticity concentration is observed in its narrow near-wake region. The increase of the gap-flow proximity interference tends to stabilize the VIV lock-in, which eventually amplifies the spanwise correlation length and weakens the streamwise vortical structures. We employ the dynamic mode decomposition procedure to characterize the space-time evolution of the primary vortex wake. △ Less

Submitted 15 January, 2018; originally announced January 2018.

Investigation on the performance of a ducted propeller in oblique flow

Authors: Qin Zhang, Rajeev K. Jaiman, Peifeng Ma, Jing Liu

Abstract: In this study, the ducted propeller has been numerically investigated under oblique flow, which is crucial and challenging for the design and safe operation of thruster driven vessel and dynamic positioning (DP) system. A Reynolds-Averaged Navier-Stokes (RANS) model has been first evaluated in the quasi-steady investigation on a single ducted propeller operating in open water condition, and then a… ▽ More In this study, the ducted propeller has been numerically investigated under oblique flow, which is crucial and challenging for the design and safe operation of thruster driven vessel and dynamic positioning (DP) system. A Reynolds-Averaged Navier-Stokes (RANS) model has been first evaluated in the quasi-steady investigation on a single ducted propeller operating in open water condition, and then a hybrid RANS/LES model is adapted for the transient sliding mesh computations. A representative test geometry considered here is a marine model thruster which is discretized with structured hexahedral cells, and the gap between the blade tip and nozzle is carefully meshed to capture the flow dynamics. The computational results are assessed by a systematic grid convergence study and compared with the available experimental data. As a part of novel contribution, multiple incidence angles from $15^\circ$ to $60^\circ$ have been analyzed with varying advance coefficients. The main emphasis has been placed on the hydrodynamic loads that act on the propeller blades and nozzle as well as their variation with different configurations. The results reveal that while the nozzle absorbs much effort from the oblique flow, the imbalance between blades at different positions is still noticeable. Such unbalance flow dynamics on the blades and the nozzle has a direct implication on the variation of thrust and torque of a marine thruster. △ Less

Submitted 9 January, 2018; originally announced January 2018.

Comments: 12 pages, 22 figures, submitted to Journal of Offshore Mechanics and Arctic Engineering

3D Common-Refinement Method for Non-Matching Meshes in Partitioned Variational Fluid-Structure Analysis

Authors: Yulong Li, Yun Zhi Law, Vaibhav Joshi, Rajeev Kumar Jaiman

Abstract: We present a three-dimensional (3D) common-refinement method for non-matching meshes between discrete non-overlapping subdomains of incompressible fluid and nonlinear hyperelastic structure. To begin, we first investigate the accuracy of common-refinement method (CRM) to satisfy traction equilibrium condition along the fluid-elastic interface with non-matching meshes. We systematically assess the… ▽ More We present a three-dimensional (3D) common-refinement method for non-matching meshes between discrete non-overlapping subdomains of incompressible fluid and nonlinear hyperelastic structure. To begin, we first investigate the accuracy of common-refinement method (CRM) to satisfy traction equilibrium condition along the fluid-elastic interface with non-matching meshes. We systematically assess the accuracy of CRM against the matching grid solution by varying grid mismatch between the fluid and solid meshes over a cylindrical tubular elastic body. We demonstrate second-order accuracy of CRM through uniform refinements of fluid and solid meshes along the interface. We then extend the error analysis to transient data transfer across non-matching meshes between fluid and solid solvers. We show that the common-refinement discretization across non-matching fluid-structure grids yields accurate transfer of the physical quantities across the fluid-solid interface. We next solve a 3D benchmark problem of a cantilevered hyperelastic plate behind a circular bluff body and verify the accuracy of coupled solutions with respect to the available solution in the literature. By varying the solid interface resolution, we generate various non-matching grid ratios and quantify the accuracy of CRM for the nonlinear structure interacting with a laminar flow. We illustrate that the CRM with the partitioned NIFC treatment is stable for low solid-to-fluid density ratio and non-matching meshes. Finally, we demonstrate the 3D parallel implementation of common-refinement with NIFC scheme for a realistic engineering problem of drilling riser undergoing complex vortex-induced vibration with strong added mass effects. △ Less

Submitted 6 November, 2017; originally announced November 2017.

Comments: 38 pages, 16 figures

A Variational Projection Scheme for Nonmatching Surface-to-Line Coupling between 3D Flexible Multibody System and Incompressible Turbulent Flow

Authors: P S Gurugubelli, R Ghoshal, V Joshi, R K Jaiman

Abstract: This paper is concerned with the partitioned iterative formulation to simulate the fluid-structure interaction of a nonlinear multibody system in an incompressible turbulent flow. The proposed formulation relies on a three-dimensional (3D) incompressible turbulent flow solver, a nonlinear monolithic elastic structural solver for constrained flexible multibody system and the nonlinear iterative for… ▽ More This paper is concerned with the partitioned iterative formulation to simulate the fluid-structure interaction of a nonlinear multibody system in an incompressible turbulent flow. The proposed formulation relies on a three-dimensional (3D) incompressible turbulent flow solver, a nonlinear monolithic elastic structural solver for constrained flexible multibody system and the nonlinear iterative force correction scheme for coupling of the turbulent fluid-flexible multibody system with nonmatching interface meshes. While the fluid equations are discretized using a stabilized Petrov-Galerkin formulation in space and the generalized-$α$ updates in time, the multibody system utilizes a discontinuous space-time Galerkin finite element method. We address two key challenges in the present formulation. Firstly, the coupling of the incompressible turbulent flow with a system of nonlinear elastic bodies described in a co-rotated frame. Secondly, the projection of the tractions and displacements across the nonmatching 3D fluid surface elements and the one-dimensional line elements for the flexible multibody system in a conservative manner. Through the nonlinear iterative correction and the conservative projection, the developed fluid-flexible multibody interaction solver is stable for problems involving strong inertial effects between the fluid-flexible multibody system and the coupled interactions among each multibody component. The accuracy of the proposed coupled finite element framework is validated against the available experimental data for a long flexible cylinder undergoing vortex-induced vibration in a uniform current flow condition. Finally, a practical application of the proposed framework is demonstrated by simulating the flow-induced vibration of a realistic offshore floating platform connected to a long riser and an elastic mooring system. △ Less

Submitted 5 November, 2017; originally announced November 2017.

On the Mechanism of Large Amplitude Flapping of Inverted Foil in a Uniform Flow

Authors: P. S. Gurugubelli, R. K. Jaiman

Abstract: An elastic foil interacting with a uniform flow with its trailing edge clamped, also known as the inverted foil, exhibits a wide range of complex self-induced flapping regimes such as large amplitude flapping (LAF), deformed and flipped flapping. Here, we perform three-dimensional numerical experiments to examine the role of vortex shedding and the vortex-vortex interaction on the LAF response at… ▽ More An elastic foil interacting with a uniform flow with its trailing edge clamped, also known as the inverted foil, exhibits a wide range of complex self-induced flapping regimes such as large amplitude flapping (LAF), deformed and flipped flapping. Here, we perform three-dimensional numerical experiments to examine the role of vortex shedding and the vortex-vortex interaction on the LAF response at Reynolds number Re=30,000. Here we investigate the dynamics of the inverted foil for a novel configuration wherein we introduce a fixed splitter plate at the trailing edge to suppress the vortex shedding from trailing edge and inhibit the interaction between the counter-rotating vortices. We find that the inhibition of the interaction has an insignificant effect on the transverse flapping amplitudes, due to a relatively weaker coupling between the counter-rotating vortices emanating from the leading edge and trailing edge. However, the inhibition of the trailing edge vortex reduces the streamwise flapping amplitude, the flapping frequency and the net strain energy of foil. To further generalize our understanding of the LAF, we next perform low-Reynolds number (Re$\in[0.1,50]$) simulations for the identical foil properties to realize the impact of vortex shedding on the large amplitude flapping. Due to the absence of vortex shedding process in the low-$Re$ regime, the inverted foil no longer exhibits the periodic flapping. However, the flexible foil still loses its stability through divergence instability to undergo a large static deformation. Finally, we introduce an analogous analytical model for the LAF based on the dynamics of an elastically mounted flat plate undergoing flow-induced pitching oscillations in a uniform stream. △ Less

Submitted 3 November, 2017; originally announced November 2017.

An adaptive variational procedure for the conservative and positivity preserving Allen-Cahn phase-field model

Authors: Vaibhav Joshi, Rajeev K. Jaiman

Abstract: We present an adaptive variational procedure for unstructured meshes to capture fluid-fluid interfaces in two-phase flows. The two phases are modeled by the phase-field finite element formulation, which involves the conservative Allen-Cahn equation coupled with the incompressible Navier-Stokes equations. The positivity preserving variational formulation is designed to maintain the bounded and stab… ▽ More We present an adaptive variational procedure for unstructured meshes to capture fluid-fluid interfaces in two-phase flows. The two phases are modeled by the phase-field finite element formulation, which involves the conservative Allen-Cahn equation coupled with the incompressible Navier-Stokes equations. The positivity preserving variational formulation is designed to maintain the bounded and stable solution of the Allen-Cahn equation. For the adaptivity procedure, we consider the residual-based error estimates for the underlying differential equations of the two-phase system. In particular, the adaptive refinement/coarsening is carried out by the newest vertex bisection algorithm by evaluating the residual error indicators based on the error estimates of the Allen-Cahn equation. The coarsening algorithm avoids the storage of the tree data structures for the hierarchical mesh, thus providing the ease of numerical implementation. The proposed nonlinear adaptive partitioned procedure aims at reducing the amount of coarsening while maintaining the convergence properties of the underlying nonlinear coupled differential equations. We investigate the adaptive phase-field finite element scheme through the spinodal decomposition in a complex curved geometry and the volume-conserved interface motion driven by the mean curvature flow for two circles in a square domain. We then assess the accuracy and efficiency of the proposed procedure by modeling the free-surface motion in a sloshing tank. The mesh adaptivity remarkably reduces the degrees of freedom and the computational cost by nearly half for similar accuracy and improves the mass conservation. Finally, we apply the adaptive numerical framework to solve the application of a dam-breaking problem with topological changes. △ Less

Submitted 6 March, 2018; v1 submitted 28 October, 2017; originally announced October 2017.

Comments: 33 pages, 20 figures, 1 table

Journal ref: J.Comp.Phy. 366 (2018) 478-504