Showing 1–8 of 8 results for author: Meena, M G

Learning the boundary-to-domain map** using Lifting Product Fourier Neural Operators for partial differential equations

Authors: Aditya Kashi, Arka Daw, Muralikrishnan Gopalakrishnan Meena, Hao Lu

Abstract: Neural operators such as the Fourier Neural Operator (FNO) have been shown to provide resolution-independent deep learning models that can learn map**s between function spaces. For example, an initial condition can be mapped to the solution of a partial differential equation (PDE) at a future time-step using a neural operator. Despite the popularity of neural operators, their use to predict solu… ▽ More Neural operators such as the Fourier Neural Operator (FNO) have been shown to provide resolution-independent deep learning models that can learn map**s between function spaces. For example, an initial condition can be mapped to the solution of a partial differential equation (PDE) at a future time-step using a neural operator. Despite the popularity of neural operators, their use to predict solution functions over a domain given only data over the boundary (such as a spatially varying Dirichlet boundary condition) remains unexplored. In this paper, we refer to such problems as boundary-to-domain problems; they have a wide range of applications in areas such as fluid mechanics, solid mechanics, heat transfer etc. We present a novel FNO-based architecture, named Lifting Product FNO (or LP-FNO) which can map arbitrary boundary functions defined on the lower-dimensional boundary to a solution in the entire domain. Specifically, two FNOs defined on the lower-dimensional boundary are lifted into the higher dimensional domain using our proposed lifting product layer. We demonstrate the efficacy and resolution independence of the proposed LP-FNO for the 2D Poisson equation. △ Less

Submitted 1 July, 2024; v1 submitted 24 June, 2024; originally announced June 2024.

Comments: Accepted by ICML 2024 AI for Science Workshop

MSC Class: 65N99; 68T07 ACM Class: I.2.1; J.2

Machine-Learned Closure of URANS for Stably Stratified Turbulence: Connecting Physical Timescales & Data Hyperparameters of Deep Time-Series Models

Authors: Muralikrishnan Gopalakrishnan Meena, Demetri Liousas, Andrew D. Simin, Aditya Kashi, Wesley H. Brewer, James J. Riley, Stephen M. de Bruyn Kops

Abstract: We develop time-series machine learning (ML) methods for closure modeling of the Unsteady Reynolds Averaged Navier Stokes (URANS) equations applied to stably stratified turbulence (SST). SST is strongly affected by fine balances between forces and becomes more anisotropic in time for decaying cases. Moreover, there is a limited understanding of the physical phenomena described by some of the terms… ▽ More We develop time-series machine learning (ML) methods for closure modeling of the Unsteady Reynolds Averaged Navier Stokes (URANS) equations applied to stably stratified turbulence (SST). SST is strongly affected by fine balances between forces and becomes more anisotropic in time for decaying cases. Moreover, there is a limited understanding of the physical phenomena described by some of the terms in the URANS equations. Rather than attempting to model each term separately, it is attractive to explore the capability of machine learning to model groups of terms, i.e., to directly model the force balances. We consider decaying SST which are homogeneous and stably stratified by a uniform density gradient, enabling dimensionality reduction. We consider two time-series ML models: Long Short-Term Memory (LSTM) and Neural Ordinary Differential Equation (NODE). Both models perform accurately and are numerically stable in a posteriori tests. Furthermore, we explore the data requirements of the ML models by extracting physically relevant timescales of the complex system. We find that the ratio of the timescales of the minimum information required by the ML models to accurately capture the dynamics of the SST corresponds to the Reynolds number of the flow. The current framework provides the backbone to explore the capability of such models to capture the dynamics of higher-dimensional complex SST flows. △ Less

Submitted 24 April, 2024; originally announced April 2024.

Insights into the interplay between hydrodynamic and acoustic fields in a turbulent combustor via community-based dimensionality reduction of vortical networks

Authors: Ankit Sahay, Muralikrishnan Gopalakrishnan Meena, R. I. Sujith

Abstract: This study examines the interplay between acoustic pressure oscillations and interactions between regions of intense vorticity in the reactive field of a turbulent combustor. These regions of intense vortical interactions are identified as vortical communities in the network space of weighted directed vortical networks constructed from two-dimensional velocity data. The vortical interactions in th… ▽ More This study examines the interplay between acoustic pressure oscillations and interactions between regions of intense vorticity in the reactive field of a turbulent combustor. These regions of intense vortical interactions are identified as vortical communities in the network space of weighted directed vortical networks constructed from two-dimensional velocity data. The vortical interactions in the high-dimensional reactive flow are condensed into a low-dimensional network representation by leveraging the inter-community strengths and the weighted community centroids. Subsequently, we show that the mean and the maximum of all inter-community interactions exhibit a strong delayed correlation with the acoustic pressure oscillations during the state of thermoacoustic instability. In contrast, during the state of combustion noise, the correlation between the acoustic pressure oscillations and the inter-community interactions is lost due to the incoherent spatiotemporal behaviour of acoustic and hydrodynamic fields in turbulent thermoacoustic system. Spatiotemporal evolution of pairs of vortical communities with the maximum inter-community interactions provides insight into explaining the critical regions detected in the reaction field during the states of intermittency and thermoacoustic instability in previous studies. We further demonstrate that the high correlations between network measures and acoustic pressure oscillations during the state of thermoacoustic instability weaken when steady air jets are introduced within the critical region to suppress the thermoacoustic oscillations. This work sheds light on the intricate relationship between the vortical interactions and the acoustic behaviour in a turbulent combustor, thus offering insights for understanding thermoacoustic instabilities in combustion systems and for unsteady flows in general fluid dynamics literature. △ Less

Submitted 21 November, 2023; originally announced November 2023.

Comments: 33 pages, 18 figures

Network broadcast analysis and control of turbulent flows

Authors: Chi-An Yeh, Muralikrishnan Gopalakrishnan Meena, Kunihiko Taira

Abstract: We present a network-based modal analysis technique that identifies key dynamical paths along which perturbations amplify over a time-varying base flow. This analysis is built upon the Katz centrality, which reveals the flow structures that can effectively spread perturbations over a time-evolving network of vortical interactions on the base flow. Motivated by the resolvent form of the Katz functi… ▽ More We present a network-based modal analysis technique that identifies key dynamical paths along which perturbations amplify over a time-varying base flow. This analysis is built upon the Katz centrality, which reveals the flow structures that can effectively spread perturbations over a time-evolving network of vortical interactions on the base flow. Motivated by the resolvent form of the Katz function, we take the singular value decomposition of the resulting communicability matrix, complementing the resolvent analysis for fluid flows. The right-singular vectors, referred to as the broadcast modes, give insights into the sensitive regions where introduced perturbations can be effectively spread and amplified over the entire fluid-flow network that evolves in time. We apply this analysis to a two-dimensional decaying isotropic turbulence. The broadcast mode reveals that vortex dipoles are important structures in spreading perturbations. By perturbing the flow with the principal broadcast mode, we demonstrate the utility of the insights gained from the present analysis to effectively modify the evolution of turbulent flows. The current network-inspired work presents a novel use of network analysis to guide flow control efforts, in particular for time-varying base flows. △ Less

Submitted 28 October, 2020; v1 submitted 29 June, 2020; originally announced June 2020.

Identifying vortical network connectors for turbulent flow modification

Authors: Muralikrishnan Gopalakrishnan Meena, Kunihiko Taira

Abstract: We introduce a network (graph) theoretic community-based framework to extract vortical structures that serve the role of connectors for the vortical interactions in two- and three-dimensional isotropic turbulence. The present framework represents the vortical interactions on a network, where the vortical elements are viewed as the nodes and the vortical interactions are regarded as edges weighted… ▽ More We introduce a network (graph) theoretic community-based framework to extract vortical structures that serve the role of connectors for the vortical interactions in two- and three-dimensional isotropic turbulence. The present framework represents the vortical interactions on a network, where the vortical elements are viewed as the nodes and the vortical interactions are regarded as edges weighted by induced velocity. We identify closely interacting vortical elements as vortical network communities through community detection algorithms. We show that the inter- and intra-community interactions can be used to decompose the governing equation for the evolution of network nodes. These community-based interactions are used to identify the communities which have the strongest and weakest interactions amongst them. These vortical communities are referred to as connector and peripheral communities, respectively. We demonstrate the influence of the network-based structures to modify the dynamics of a collection of discrete point vortices. Taking advantage of the strong inter-community interactions, connector community can significantly modify the collective dynamics of vortices through the application of multiple impulse perturbations. We then apply the community-based framework to extract influential structures in isotropic turbulence. The connector and peripheral communities extracted from turbulent flows resemble shear-layer and vortex-core like structures, respectively. The influence of the connector structures on the flow field and their neighboring vortical structures is analyzed by adding impulse perturbations to the connectors in direct numerical simulations. The findings are compared with the cases of perturbing the strongest vortex tube and shear-layer regions. We find that perturbing the connector structures enhances local turbulent mixing beyond what are achieved by the other cases. △ Less

Submitted 22 May, 2020; v1 submitted 5 May, 2020; originally announced May 2020.

Comments: Update: Corrected equation (2.12)

Journal ref: J. Fluid Mech. 915 (2021) A10

Randomized methods to characterize large-scale vortical flow network

Authors: Zhe Bai, N. Benjamin Erichson, Muralikrishnan Gopalakrishnan Meena, Kunihiko Taira, Steven L. Brunton

Abstract: We demonstrate the effective use of randomized methods for linear algebra to perform network-based analysis of complex vortical flows. Network theoretic approaches can reveal the connectivity structures among a set of vortical elements and analyze their collective dynamics. These approaches have recently been generalized to analyze high-dimensional turbulent flows, for which network computations c… ▽ More We demonstrate the effective use of randomized methods for linear algebra to perform network-based analysis of complex vortical flows. Network theoretic approaches can reveal the connectivity structures among a set of vortical elements and analyze their collective dynamics. These approaches have recently been generalized to analyze high-dimensional turbulent flows, for which network computations can become prohibitively expensive. In this work, we propose efficient methods to approximate network quantities, such as the leading eigendecomposition of the adjacency matrix, using randomized methods. Specifically, we use the Nyström method to approximate the leading eigenvalues and eigenvectors, achieving significant computational savings and reduced memory requirements. The effectiveness of the proposed technique is demonstrated on two high-dimensional flow fields: two-dimensional flow past an airfoil and two-dimensional turbulence. We find that quasi-uniform column sampling outperforms uniform column sampling, while both feature the same computational complexity. △ Less

Submitted 2 September, 2019; originally announced September 2019.

Comments: 18 pages, 8 figures

Network community-based model reduction for vortical flows

Authors: Muralikrishnan Gopalakrishnan Meena, Aditya G. Nair, Kunihiko Taira

Abstract: A network community-based reduced-order model is developed to capture key interactions amongst coherent structures in high-dimensional unsteady vortical flows. The present approach is data-inspired and founded on network-theoretic techniques to identify important vortical communities that are comprised of vortical elements that share similar dynamical behavior. The overall interaction-based physic… ▽ More A network community-based reduced-order model is developed to capture key interactions amongst coherent structures in high-dimensional unsteady vortical flows. The present approach is data-inspired and founded on network-theoretic techniques to identify important vortical communities that are comprised of vortical elements that share similar dynamical behavior. The overall interaction-based physics of the high-dimensional flow field is distilled into the vortical community centroids, considerably reducing the system dimension. Taking advantage of these vortical interactions, the proposed methodology is applied to formulate reduced-order models for the inter-community dynamics of vortical flows, and predict lift and drag forces on bodies in wake flows. We demonstrate the capabilities of these models by accurately capturing the macroscopic dynamics of a collection of discrete point vortices, and the complex unsteady aerodynamic forces on a circular cylinder and an airfoil with a Gurney flap. The present formulation is found to be robust against simulated experimental noise and turbulence due to its integrating nature of the system reduction. △ Less

Submitted 2 May, 2018; originally announced May 2018.

Journal ref: Phys. Rev. E 97, 063103 (2018)

Airfoil wake modification with Gurney flap at Low-Reynolds number

Authors: Muralikrishnan Gopalakrishnan Meena, Kunihiko Taira, Keisuke Asai

Abstract: The complex wake modifications produced by a Gurney flap on symmetric NACA airfoils at low Reynolds number are investigated. Two-dimensional incompressible flows over NACA 0000 (flat plate), 0006, 0012 and 0018 airfoils at a Reynolds number of $Re = 1000$ are analyzed numerically to examine the flow modifications generated by the flaps for achieving lift enhancement. While high lift can be attaine… ▽ More The complex wake modifications produced by a Gurney flap on symmetric NACA airfoils at low Reynolds number are investigated. Two-dimensional incompressible flows over NACA 0000 (flat plate), 0006, 0012 and 0018 airfoils at a Reynolds number of $Re = 1000$ are analyzed numerically to examine the flow modifications generated by the flaps for achieving lift enhancement. While high lift can be attained by the Gurney flap on airfoils at high angles of attack, highly unsteady nature of the aerodynamic forces are also observed. Analysis of the wake structures along with the lift spectra reveals four characteristic wake modes (steady, 2S, P and 2P), influencing the aerodynamic performance. The effects of the flap over wide range of angles of attack and flap heights are considered to identify the occurrence of these wake modes, and are encapsulated in a wake classification diagram. Companion three-dimensional simulations are also performed to examine the influence of three-dimensionality on the wake regimes. The spanwise instabilities that appear for higher angles of attack are found to suppress the emergence of the 2P mode. The use of the wake classification diagram as a guidance for Gurney flap selection at different operating conditions to achieve the required aerodynamic performance is discussed. △ Less

Submitted 25 October, 2017; v1 submitted 28 August, 2017; originally announced August 2017.

Journal ref: AIAA Journal, Vol. 56, No. 4 (2018)