skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

This content will become publicly available on November 12, 2020

Title: Nonintrusive reduced order modeling framework for quasigeostrophic turbulence

Abstract

In this study, we present a nonintrusive reduced order modeling (ROM) framework for large-scale quasistationary systems. The framework proposed in this work exploits the time series prediction capability of long short-term memory (LSTM) recurrent neural network architecture such that (1) in the training phase, the LSTM model is trained on the modal coefficients extracted from the high-resolution data snapshots using proper orthogonal decomposition (POD) transform, and (2) in the testing phase, the trained model predicts the modal coefficients for the total time recursively based on the initial time history. Hence, no prior information about the underlying governing equations is required to generate the ROM. To illustrate the predictive performance of the proposed framework, the mean flow fields and time series response of the field values are reconstructed from the predicted modal coefficients by using an inverse POD transform. As a representative benchmark test case, we consider a two-dimensional quasigeostrophic ocean circulation model which, in general, displays an enormous range of fluctuating spatial and temporal scales. We first demonstrate that the conventional Galerkin projection-based reduced order modeling of such systems requires a high number of POD modes to obtain a stable flow physics. In addition, ROM-Galerkin projection (ROM-GP) does not seemmore » to capture the intermittent bursts appearing in the dynamics of the first few most energetic modes. However, the proposed nonintrusive ROM framework based on LSTM (ROM-LSTM) yields a stable solution even for a small number of POD modes. We also observe that the ROM-LSTM model is able to capture quasiperiodic intermittent bursts accurately, and yields a stable and accurate mean flow dynamics using the time history of a few previous time states, denoted as the lookback time window in this paper. We show several features of ROM-LSTM framework such as significantly higher accuracy than ROM-GP, and faster performance using larger time step size. Throughout the paper, we demonstrate our findings in terms of time series evolution of the field values and mean flow patterns, which suggest that the proposed fully nonintrusive ROM framework is robust and capable of predicting chaotic nonlinear fluid flows in an extremely efficient way compared to the conventional projection-based ROM framework.« less

Authors:
ORCiD logo [1];  [1]; ORCiD logo [1];  [2]; ORCiD logo [3]
  1. Oklahoma State Univ., Stillwater, OK (United States)
  2. Norwegian Univ. of Science and Technology, Trondheim (Norway)
  3. Virginia Polytechnic Inst. and State Univ. (Virginia Tech), Blacksburg, VA (United States)
Publication Date:
Research Org.:
Oklahoma State Univ., Stillwater, OK (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR) (SC-21)
OSTI Identifier:
1593558
Grant/Contract Number:  
SC0019290
Resource Type:
Accepted Manuscript
Journal Name:
Physical Review E
Additional Journal Information:
Journal Volume: 100; Journal Issue: 5; Journal ID: ISSN 2470-0045
Publisher:
American Physical Society (APS)
Country of Publication:
United States
Language:
English
Subject:
42 ENGINEERING

Citation Formats

Rahman, Sk. M., Pawar, S., San, O., Rasheed, A., and Iliescu, T. Nonintrusive reduced order modeling framework for quasigeostrophic turbulence. United States: N. p., 2019. Web. doi:10.1103/PhysRevE.100.053306.
Rahman, Sk. M., Pawar, S., San, O., Rasheed, A., & Iliescu, T. Nonintrusive reduced order modeling framework for quasigeostrophic turbulence. United States. doi:10.1103/PhysRevE.100.053306.
Rahman, Sk. M., Pawar, S., San, O., Rasheed, A., and Iliescu, T. Tue . "Nonintrusive reduced order modeling framework for quasigeostrophic turbulence". United States. doi:10.1103/PhysRevE.100.053306.
@article{osti_1593558,
title = {Nonintrusive reduced order modeling framework for quasigeostrophic turbulence},
author = {Rahman, Sk. M. and Pawar, S. and San, O. and Rasheed, A. and Iliescu, T.},
abstractNote = {In this study, we present a nonintrusive reduced order modeling (ROM) framework for large-scale quasistationary systems. The framework proposed in this work exploits the time series prediction capability of long short-term memory (LSTM) recurrent neural network architecture such that (1) in the training phase, the LSTM model is trained on the modal coefficients extracted from the high-resolution data snapshots using proper orthogonal decomposition (POD) transform, and (2) in the testing phase, the trained model predicts the modal coefficients for the total time recursively based on the initial time history. Hence, no prior information about the underlying governing equations is required to generate the ROM. To illustrate the predictive performance of the proposed framework, the mean flow fields and time series response of the field values are reconstructed from the predicted modal coefficients by using an inverse POD transform. As a representative benchmark test case, we consider a two-dimensional quasigeostrophic ocean circulation model which, in general, displays an enormous range of fluctuating spatial and temporal scales. We first demonstrate that the conventional Galerkin projection-based reduced order modeling of such systems requires a high number of POD modes to obtain a stable flow physics. In addition, ROM-Galerkin projection (ROM-GP) does not seem to capture the intermittent bursts appearing in the dynamics of the first few most energetic modes. However, the proposed nonintrusive ROM framework based on LSTM (ROM-LSTM) yields a stable solution even for a small number of POD modes. We also observe that the ROM-LSTM model is able to capture quasiperiodic intermittent bursts accurately, and yields a stable and accurate mean flow dynamics using the time history of a few previous time states, denoted as the lookback time window in this paper. We show several features of ROM-LSTM framework such as significantly higher accuracy than ROM-GP, and faster performance using larger time step size. Throughout the paper, we demonstrate our findings in terms of time series evolution of the field values and mean flow patterns, which suggest that the proposed fully nonintrusive ROM framework is robust and capable of predicting chaotic nonlinear fluid flows in an extremely efficient way compared to the conventional projection-based ROM framework.},
doi = {10.1103/PhysRevE.100.053306},
journal = {Physical Review E},
number = 5,
volume = 100,
place = {United States},
year = {2019},
month = {11}
}

Journal Article:
Free Publicly Available Full Text
This content will become publicly available on November 12, 2020
Publisher's Version of Record

Save / Share:

Works referenced in this record:

A Reduced-Order Method for Simulation and Control of Fluid Flows
journal, July 1998


Learning long-term dependencies with gradient descent is difficult
journal, March 1994

  • Bengio, Y.; Simard, P.; Frasconi, P.
  • IEEE Transactions on Neural Networks, Vol. 5, Issue 2
  • DOI: 10.1109/72.279181

Data-assisted reduced-order modeling of extreme events in complex dynamical systems
journal, May 2018


Spatio-temporal symmetries and bifurcations via bi-orthogonal decompositions
journal, June 1992

  • Aubry, N.; Guyonnet, R.; Lima, R.
  • Journal of Nonlinear Science, Vol. 2, Issue 2
  • DOI: 10.1007/BF02429855

A reduced-order model for heat transfer in multiphase flow and practical aspects of the proper orthogonal decomposition
journal, August 2012


Reduced Basis Methods for Parameterized Partial Differential Equations with Stochastic Influences Using the Karhunen--Loève Expansion
journal, January 2013

  • Haasdonk, Bernard; Urban, Karsten; Wieland, Bernhard
  • SIAM/ASA Journal on Uncertainty Quantification, Vol. 1, Issue 1
  • DOI: 10.1137/120876745

Deep learning
journal, May 2015

  • LeCun, Yann; Bengio, Yoshua; Hinton, Geoffrey
  • Nature, Vol. 521, Issue 7553
  • DOI: 10.1038/nature14539

A neural network approach for the blind deconvolution of turbulent flows
journal, October 2017


On the stability and extension of reduced-order Galerkin models in incompressible flows: A numerical study of vortex shedding
journal, June 2009

  • Akhtar, Imran; Nayfeh, Ali H.; Ribbens, Calvin J.
  • Theoretical and Computational Fluid Dynamics, Vol. 23, Issue 3
  • DOI: 10.1007/s00162-009-0112-y

Hybrid forecasting of chaotic processes: Using machine learning in conjunction with a knowledge-based model
journal, April 2018

  • Pathak, Jaideep; Wikner, Alexander; Fussell, Rebeckah
  • Chaos: An Interdisciplinary Journal of Nonlinear Science, Vol. 28, Issue 4
  • DOI: 10.1063/1.5028373

Data-driven forecasting of high-dimensional chaotic systems with long short-term memory networks
journal, May 2018

  • Vlachas, Pantelis R.; Byeon, Wonmin; Wan, Zhong Y.
  • Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 474, Issue 2213
  • DOI: 10.1098/rspa.2017.0844

A Short Review on Model Order Reduction Based on Proper Generalized Decomposition
journal, October 2011

  • Chinesta, Francisco; Ladeveze, Pierre; Cueto, Elías
  • Archives of Computational Methods in Engineering, Vol. 18, Issue 4
  • DOI: 10.1007/s11831-011-9064-7

Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
journal, February 2019


Unsteady Fluid Mechanics Applications of Neural Networks
journal, January 1997

  • Faller, William E.; Schreck, Scott J.
  • Journal of Aircraft, Vol. 34, Issue 1
  • DOI: 10.2514/2.2134

Climate change, the Hurst phenomenon, and hydrological statistics
journal, February 2003


Identification strategies for model-based control
journal, July 2013


Turbulence Modeling in the Age of Data
journal, January 2019


An improved algorithm for the shallow water equations model reduction: Dynamic Mode Decomposition vs POD: AN IMPROVED ALGORITHM FOR DYNAMIC MODE DECOMPOSITION VS POD
journal, April 2015

  • Bistrian, D. A.; Navon, I. M.
  • International Journal for Numerical Methods in Fluids, Vol. 78, Issue 9
  • DOI: 10.1002/fld.4029

Application of neural networks to turbulence control for drag reduction
journal, June 1997

  • Lee, Changhoon; Kim, John; Babcock, David
  • Physics of Fluids, Vol. 9, Issue 6
  • DOI: 10.1063/1.869290

LSTM: A Search Space Odyssey
journal, October 2017

  • Greff, Klaus; Srivastava, Rupesh K.; Koutnik, Jan
  • IEEE Transactions on Neural Networks and Learning Systems, Vol. 28, Issue 10
  • DOI: 10.1109/TNNLS.2016.2582924

A hierarchy of low-dimensional models for the transient and post-transient cylinder wake
journal, December 2003

  • Noack, Bernd R.; Afanasiev, Konstantin; MorzyŃSki, Marek
  • Journal of Fluid Mechanics, Vol. 497
  • DOI: 10.1017/S0022112003006694

Model Reduction for Flow Analysis and Control
journal, January 2017


POD and CVT-based reduced-order modeling of Navier–Stokes flows
journal, December 2006

  • Burkardt, John; Gunzburger, Max; Lee, Hyung-Chun
  • Computer Methods in Applied Mechanics and Engineering, Vol. 196, Issue 1-3
  • DOI: 10.1016/j.cma.2006.04.004

A dynamic closure modeling framework for model order reduction of geophysical flows
journal, April 2019

  • Rahman, Sk. M.; Ahmed, S. E.; San, O.
  • Physics of Fluids, Vol. 31, Issue 4
  • DOI: 10.1063/1.5093355

Approximate deconvolution reduced order modeling
journal, January 2017

  • Xie, Xuping; Wells, David; Wang, Zhu
  • Computer Methods in Applied Mechanics and Engineering, Vol. 313
  • DOI: 10.1016/j.cma.2016.10.005

On Low-Dimensional Galerkin Models for Fluid Flow
journal, June 2000


Reduced order models based on local POD plus Galerkin projection
journal, April 2010


An artificial neural network framework for reduced order modeling of transient flows
journal, October 2019

  • San, Omer; Maulik, Romit; Ahmed, Mansoor
  • Communications in Nonlinear Science and Numerical Simulation, Vol. 77
  • DOI: 10.1016/j.cnsns.2019.04.025

Total variation diminishing Runge-Kutta schemes
journal, January 1998

  • Gottlieb, Sigal; Shu, Chi-Wang
  • Mathematics of Computation of the American Mathematical Society, Vol. 67, Issue 221
  • DOI: 10.1090/S0025-5718-98-00913-2

Non-intrusive reduced order modeling of unsteady flows using artificial neural networks with application to a combustion problem
journal, May 2019


Dispersive–Dissipative Eddy Parameterization in a Barotropic Model
journal, August 2001


Long Short-Term Memory
journal, November 1997


A reduced-order approach to four-dimensional variational data assimilation using proper orthogonal decomposition
journal, January 2007

  • Cao, Yanhua; Zhu, Jiang; Navon, I. M.
  • International Journal for Numerical Methods in Fluids, Vol. 53, Issue 10
  • DOI: 10.1002/fld.1365

A Survey of Projection-Based Model Reduction Methods for Parametric Dynamical Systems
journal, January 2015

  • Benner, Peter; Gugercin, Serkan; Willcox, Karen
  • SIAM Review, Vol. 57, Issue 4
  • DOI: 10.1137/130932715

A spectral viscosity method for correcting the long-term behavior of POD models
journal, February 2004


Galerkin Proper Orthogonal Decomposition Methods for a General Equation in Fluid Dynamics
journal, January 2002


Spectral proper orthogonal decomposition
journal, March 2016

  • Sieber, Moritz; Paschereit, C. Oliver; Oberleithner, Kilian
  • Journal of Fluid Mechanics, Vol. 792
  • DOI: 10.1017/jfm.2016.103

Some Recent Developments in Turbulence Closure Modeling
journal, January 2018


Observing the Ocean in the 1990s
journal, October 1982

  • Munk, W.; Wunsch, C.
  • Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 307, Issue 1499
  • DOI: 10.1098/rsta.1982.0120

A proper orthogonal decomposition method for nonlinear flows with deforming meshes
journal, December 2014


Proper orthogonal decomposition and low-dimensional models for driven cavity flows
journal, July 1998

  • Cazemier, W.; Verstappen, R. W. C. P.; Veldman, A. E. P.
  • Physics of Fluids, Vol. 10, Issue 7
  • DOI: 10.1063/1.869686

Proper orthogonal decomposition closure models for turbulent flows: A numerical comparison
journal, September 2012

  • Wang, Zhu; Akhtar, Imran; Borggaard, Jeff
  • Computer Methods in Applied Mechanics and Engineering, Vol. 237-240
  • DOI: 10.1016/j.cma.2012.04.015

Robustness of the rescaled range R/S in the measurement of noncyclic long run statistical dependence
journal, October 1969


Neural network closures for nonlinear model order reduction
journal, January 2018


Noah, Joseph, and Operational Hydrology
journal, October 1968


Wind-driven ocean circulation and equilibrium statistical mechanics
journal, August 1989


The End of Moore's Law: A New Beginning for Information Technology
journal, March 2017

  • Theis, Thomas N.; Wong, H. -S. Philip
  • Computing in Science & Engineering, Vol. 19, Issue 2
  • DOI: 10.1109/MCSE.2017.29

Four-Gyre Circulation in a Barotropic Model with Double-Gyre Wind Forcing
journal, June 2000


Closed-Loop Turbulence Control: Progress and Challenges
journal, August 2015

  • Brunton, Steven L.; Noack, Bernd R.
  • Applied Mechanics Reviews, Vol. 67, Issue 5
  • DOI: 10.1115/1.4031175

Data-driven deconvolution for large eddy simulations of Kraichnan turbulence
journal, December 2018

  • Maulik, R.; San, O.; Rasheed, A.
  • Physics of Fluids, Vol. 30, Issue 12
  • DOI: 10.1063/1.5079582

Fifty Years of Moore's Law
journal, May 2011


The dynamics of coherent structures in the wall region of a turbulent boundary layer
journal, July 1988


On the need for a nonlinear subscale turbulence term in POD models as exemplified for a high-Reynolds-number flow over an Ahmed body
journal, April 2014

  • Östh, Jan; Noack, Bernd R.; Krajnović, Siniša
  • Journal of Fluid Mechanics, Vol. 747
  • DOI: 10.1017/jfm.2014.168

Machine Learning for Fluid Mechanics
journal, September 2019


Emergence of Inertial Gyres in a Two-Layer Quasigeostrophic Ocean Model
journal, March 1998


An Example of Eddy-Induced Ocean Circulation
journal, July 1980


Reduced-order modeling: new approaches for computational physics
journal, February 2004


Deep learning algorithm for data-driven simulation of noisy dynamical system
journal, January 2019


New approach for automatic classification of Alzheimer's disease, mild cognitive impairment and healthy brain magnetic resonance images
journal, January 2014

  • Boukadoum, Mounir; Lahmiri, Salim
  • Healthcare Technology Letters, Vol. 1, Issue 1
  • DOI: 10.1049/htl.2013.0022

Efficient non-linear model reduction via a least-squares Petrov-Galerkin projection and compressive tensor approximations
journal, October 2010

  • Carlberg, Kevin; Bou-Mosleh, Charbel; Farhat, Charbel
  • International Journal for Numerical Methods in Engineering, Vol. 86, Issue 2
  • DOI: 10.1002/nme.3050

Approximate deconvolution large eddy simulation of a barotropic ocean circulation model
journal, January 2011


Model reduction for compressible flows using POD and Galerkin projection
journal, February 2004

  • Rowley, Clarence W.; Colonius, Tim; Murray, Richard M.
  • Physica D: Nonlinear Phenomena, Vol. 189, Issue 1-2
  • DOI: 10.1016/j.physd.2003.03.001

The chips are down for Moore’s law
journal, February 2016


The Proper Orthogonal Decomposition in the Analysis of Turbulent Flows
journal, January 1993


Wind-Driven Currents in a Baroclinic Ocean; with Application to the Equatorial Currents of the Eastern Pacific
journal, November 1947

  • Sverdrup, H. U.
  • Proceedings of the National Academy of Sciences, Vol. 33, Issue 11
  • DOI: 10.1073/pnas.33.11.318

A Hybrid Approach for Model Order Reduction of Barotropic Quasi-Geostrophic Turbulence
journal, October 2018


Stability Properties of POD-Galerkin Approximations for the Compressible Navier-Stokes Equations
journal, March 2000

  • Iollo, A.; Lanteri, S.; Désidéri, J. -A.
  • Theoretical and Computational Fluid Dynamics, Vol. 13, Issue 6
  • DOI: 10.1007/s001620050119

An Efficient Coarse grid Projection Method for Quasigeostrophic Models of Large-Scale Ocean Circulation
journal, January 2013


A stabilized proper orthogonal decomposition reduced-order model for large scale quasigeostrophic ocean circulation
journal, May 2015


Towards a topological–geometrical theory of group equivariant non-expansive operators for data analysis and machine learning
journal, September 2019

  • Bergomi, Mattia G.; Frosini, Patrizio; Giorgi, Daniela
  • Nature Machine Intelligence, Vol. 1, Issue 9
  • DOI: 10.1038/s42256-019-0087-3

Model identification of reduced order fluid dynamics systems using deep learning: Model identification in fluid dynamics using deep learning
journal, August 2017

  • Wang, Z.; Xiao, D.; Fang, F.
  • International Journal for Numerical Methods in Fluids, Vol. 86, Issue 4
  • DOI: 10.1002/fld.4416

Turbulence and the dynamics of coherent structures. I. Coherent structures
journal, January 1987

  • Sirovich, Lawrence
  • Quarterly of Applied Mathematics, Vol. 45, Issue 3
  • DOI: 10.1090/qam/910462

Markov processes, Hurst exponents, and nonlinear diffusion equations: With application to finance
journal, September 2006

  • Bassler, Kevin E.; Gunaratne, Gemunu H.; McCauley, Joseph L.
  • Physica A: Statistical Mechanics and its Applications, Vol. 369, Issue 2
  • DOI: 10.1016/j.physa.2006.01.081

Time-dependent Hurst exponent in financial time series
journal, December 2004

  • Carbone, A.; Castelli, G.; Stanley, H. E.
  • Physica A: Statistical Mechanics and its Applications, Vol. 344, Issue 1-2
  • DOI: 10.1016/j.physa.2004.06.130

Deep learning in fluid dynamics
journal, January 2017


Inertial gyres in decaying and forced geostrophic turbulence
journal, November 1992


Modeling Mesoscale Turbulence in the Barotropic Double-Gyre Circulation
journal, November 2003


Dynamics and Control of High-Reynolds-Number flow over open Cavities
journal, January 2006