Multi-fidelity information fusion with concatenated neural networks
Abstract
Recently, computational modeling has shifted towards the use of statistical inference, deep learning, and other data-driven modeling frameworks. Although this shift in modeling holds promise in many applications like design optimization and real-time control by lowering the computational burden, training deep learning models needs a huge amount of data. This big data is not always available for scientific problems and leads to poorly generalizable data-driven models. This gap can be furnished by leveraging information from physics-based models. Exploiting prior knowledge about the problem at hand, this study puts forth a physics-guided machine learning (PGML) approach to build more tailored, effective, and efficient surrogate models. For our analysis, without losing its generalizability and modularity, we focus on the development of predictive models for laminar and turbulent boundary layer flows. In particular, we combine the self-similarity solution and power-law velocity profile (low-fidelity models) with the noisy data obtained either from experiments or computational fluid dynamics simulations (high-fidelity models) through a concatenated neural network. We illustrate how the knowledge from these simplified models results in reducing uncertainties associated with deep learning models applied to boundary layer flow prediction problems. The proposed multi-fidelity information fusion framework produces physically consistent models that attempt to achievemore »
- Authors:
-
- Oklahoma State Univ., Stillwater, OK (United States)
- Univ. of Oklahoma, Norman, OK (United States)
- Norwegian Univ. of Science and Technology, Trondheim (Norway); SINTEF Digital, Trondheim (Norway)
- Publication Date:
- Research Org.:
- Oklahoma State Univ., Stillwater, OK (United States)
- Sponsoring Org.:
- USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
- OSTI Identifier:
- 1904349
- Grant/Contract Number:
- SC0019290
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Scientific Reports
- Additional Journal Information:
- Journal Volume: 12; Journal Issue: 1; Journal ID: ISSN 2045-2322
- Publisher:
- Nature Publishing Group
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 42 ENGINEERING; aerospace engineering; applied mathematics; computational science; mechanical engineering
Citation Formats
Pawar, Suraj, San, Omer, Vedula, Prakash, Rasheed, Adil, and Kvamsdal, Trond. Multi-fidelity information fusion with concatenated neural networks. United States: N. p., 2022.
Web. doi:10.1038/s41598-022-09938-8.
Pawar, Suraj, San, Omer, Vedula, Prakash, Rasheed, Adil, & Kvamsdal, Trond. Multi-fidelity information fusion with concatenated neural networks. United States. https://doi.org/10.1038/s41598-022-09938-8
Pawar, Suraj, San, Omer, Vedula, Prakash, Rasheed, Adil, and Kvamsdal, Trond. Thu .
"Multi-fidelity information fusion with concatenated neural networks". United States. https://doi.org/10.1038/s41598-022-09938-8. https://www.osti.gov/servlets/purl/1904349.
@article{osti_1904349,
title = {Multi-fidelity information fusion with concatenated neural networks},
author = {Pawar, Suraj and San, Omer and Vedula, Prakash and Rasheed, Adil and Kvamsdal, Trond},
abstractNote = {Recently, computational modeling has shifted towards the use of statistical inference, deep learning, and other data-driven modeling frameworks. Although this shift in modeling holds promise in many applications like design optimization and real-time control by lowering the computational burden, training deep learning models needs a huge amount of data. This big data is not always available for scientific problems and leads to poorly generalizable data-driven models. This gap can be furnished by leveraging information from physics-based models. Exploiting prior knowledge about the problem at hand, this study puts forth a physics-guided machine learning (PGML) approach to build more tailored, effective, and efficient surrogate models. For our analysis, without losing its generalizability and modularity, we focus on the development of predictive models for laminar and turbulent boundary layer flows. In particular, we combine the self-similarity solution and power-law velocity profile (low-fidelity models) with the noisy data obtained either from experiments or computational fluid dynamics simulations (high-fidelity models) through a concatenated neural network. We illustrate how the knowledge from these simplified models results in reducing uncertainties associated with deep learning models applied to boundary layer flow prediction problems. The proposed multi-fidelity information fusion framework produces physically consistent models that attempt to achieve better generalization than data-driven models obtained purely based on data. While we demonstrate our framework for a problem relevant to fluid mechanics, its workflow and principles can be adopted for many scientific problems where empirical, analytical, or simplified models are prevalent. In line with grand demands in novel PGML principles, this work builds a bridge between extensive physics-based theories and data-driven modeling paradigms and paves the way for using hybrid physics and machine learning modeling approaches for next-generation digital twin technologies.},
doi = {10.1038/s41598-022-09938-8},
journal = {Scientific Reports},
number = 1,
volume = 12,
place = {United States},
year = {Thu Apr 07 00:00:00 EDT 2022},
month = {Thu Apr 07 00:00:00 EDT 2022}
}
Works referenced in this record:
Adversarial super-resolution of climatological wind and solar data
journal, July 2020
- Stengel, Karen; Glaws, Andrew; Hettinger, Dylan
- Proceedings of the National Academy of Sciences, Vol. 117, Issue 29
Physics-informed machine learning: case studies for weather and climate modelling
journal, February 2021
- Kashinath, K.; Mustafa, M.; Albert, A.
- Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 379, Issue 2194
Predicting clustered weather patterns: A test case for applications of convolutional neural networks to spatio-temporal climate data
journal, January 2020
- Chattopadhyay, Ashesh; Hassanzadeh, Pedram; Pasha, Saba
- Scientific Reports, Vol. 10, Issue 1
Physics-informed machine learning
journal, May 2021
- Karniadakis, George Em; Kevrekidis, Ioannis G.; Lu, Lu
- Nature Reviews Physics, Vol. 3, Issue 6
Solution of the implicitly discretised fluid flow equations by operator-splitting
journal, January 1986
- Issa, R. I.
- Journal of Computational Physics, Vol. 62, Issue 1
Panel Methods in Computational Fluid Dynamics
journal, January 1990
- Hess, J. L.
- Annual Review of Fluid Mechanics, Vol. 22, Issue 1
Deep learning and process understanding for data-driven Earth system science
journal, February 2019
- Reichstein, Markus; Camps-Valls, Gustau; Stevens, Bjorn
- Nature, Vol. 566, Issue 7743
Evaluation, Tuning, and Interpretation of Neural Networks for Working with Images in Meteorological Applications
journal, December 2020
- Ebert-Uphoff, Imme; Hilburn, Kyle
- Bulletin of the American Meteorological Society, Vol. 101, Issue 12
Subgrid modelling for two-dimensional turbulence using neural networks
journal, November 2018
- Maulik, R.; San, O.; Rasheed, A.
- Journal of Fluid Mechanics, Vol. 858
Wall-bounded turbulence
journal, September 2013
- Smits, Alexander J.; Marusic, Ivan
- Physics Today, Vol. 66, Issue 9
DIRECT NUMERICAL SIMULATION: A Tool in Turbulence Research
journal, January 1998
- Moin, Parviz; Mahesh, Krishnan
- Annual Review of Fluid Mechanics, Vol. 30, Issue 1
Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
journal, February 2019
- Raissi, M.; Perdikaris, P.; Karniadakis, G. E.
- Journal of Computational Physics, Vol. 378
Enforcing Analytic Constraints in Neural Networks Emulating Physical Systems
journal, March 2021
- Beucler, Tom; Pritchard, Michael; Rasp, Stephan
- Physical Review Letters, Vol. 126, Issue 9
Hybrid analysis and modeling, eclecticism, and multifidelity computing toward digital twin revolution
journal, May 2021
- San, Omer; Rasheed, Adil; Kvamsdal, Trond
- GAMM-Mitteilungen, Vol. 44, Issue 2
Turbulence Modeling in the Age of Data
journal, January 2019
- Duraisamy, Karthik; Iaccarino, Gianluca; Xiao, Heng
- Annual Review of Fluid Mechanics, Vol. 51, Issue 1
Grand challenges in the science of wind energy
journal, October 2019
- Veers, Paul; Dykes, Katherine; Lantz, Eric
- Science, Vol. 366, Issue 6464
Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
journal, October 2016
- Ling, Julia; Kurzawski, Andrew; Templeton, Jeremy
- Journal of Fluid Mechanics, Vol. 807
Ensemble Methods in Machine Learning
book, January 2000
- Dietterich, Thomas G.
- Multiple Classifier Systems
Machine Learning for Fluid Mechanics
journal, September 2019
- Brunton, Steven L.; Noack, Bernd R.; Koumoutsakos, Petros
- Annual Review of Fluid Mechanics, Vol. 52, Issue 1
A general approach to get series solution of non-similarity boundary-layer flows
journal, May 2009
- Liao, Shi-Jun
- Communications in Nonlinear Science and Numerical Simulation, Vol. 14, Issue 5
Effect of Roughness on Wall-Bounded Turbulence
journal, January 2004
- Bhaganagar, Kiran; Kim, John; Coleman, Gary
- Flow, Turbulence and Combustion formerly `Applied Scientific Research', Vol. 72, Issue 2-4
A Single Formula for the “Law of the Wall”
journal, September 1961
- Spalding, D. B.
- Journal of Applied Mechanics, Vol. 28, Issue 3
Data‐Driven Equation Discovery of Ocean Mesoscale Closures
journal, August 2020
- Zanna, Laure; Bolton, Thomas
- Geophysical Research Letters, Vol. 47, Issue 17
Physics guided machine learning using simplified theories
journal, January 2021
- Pawar, Suraj; San, Omer; Aksoylu, Burak
- Physics of Fluids, Vol. 33, Issue 1
Modeling the dynamics of PDE systems with physics-constrained deep auto-regressive networks
journal, February 2020
- Geneva, Nicholas; Zabaras, Nicholas
- Journal of Computational Physics, Vol. 403
Recent advances and applications of machine learning in solid-state materials science
journal, August 2019
- Schmidt, Jonathan; Marques, Mário R. G.; Botti, Silvana
- npj Computational Materials, Vol. 5, Issue 1
Making the Black Box More Transparent: Understanding the Physical Implications of Machine Learning
journal, November 2019
- McGovern, Amy; Lagerquist, Ryan; John Gagne, David
- Bulletin of the American Meteorological Society, Vol. 100, Issue 11
Machine Learning for Model Error Inference and Correction
journal, November 2020
- Bonavita, Massimo; Laloyaux, Patrick
- Journal of Advances in Modeling Earth Systems, Vol. 12, Issue 12
A survey of modelling methods for high-fidelity wind farm simulations using large eddy simulation
journal, March 2017
- Breton, S. -P.; Sumner, J.; Sørensen, J. N.
- Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 375, Issue 2091
A Comprehensive Deep Learning-Based Approach to Reduced Order Modeling of Nonlinear Time-Dependent Parametrized PDEs
journal, April 2021
- Fresca, Stefania; Dede’, Luca; Manzoni, Andrea
- Journal of Scientific Computing, Vol. 87, Issue 2
Matplotlib: A 2D Graphics Environment
journal, January 2007
- Hunter, John D.
- Computing in Science & Engineering, Vol. 9, Issue 3
Cascades in Wall-Bounded Turbulence
journal, January 2012
- Jiménez, Javier
- Annual Review of Fluid Mechanics, Vol. 44, Issue 1
Direct numerical simulation of turbulence in a nominally zero-pressure-gradient flat-plate boundary layer
journal, July 2009
- Wu, Xiaohua; Moin, Parviz
- Journal of Fluid Mechanics, Vol. 630
Review and evaluation of wake loss models for wind energy applications
journal, September 2018
- Archer, Cristina L.; Vasel-Be-Hagh, Ahmadreza; Yan, Chi
- Applied Energy, Vol. 226
Enforcing statistical constraints in generative adversarial networks for modeling chaotic dynamical systems
journal, April 2020
- Wu, Jin-Long; Kashinath, Karthik; Albert, Adrian
- Journal of Computational Physics, Vol. 406