Title: Model fusion with physics-guided machine learning: Projection-based reduced-order modeling

Abstract

The unprecedented amount of data generated from experiments, field observations, and large-scale numerical simulations at a wide range of spatiotemporal scales has enabled the rapid advancement of data-driven and especially deep learning models in the field of fluid mechanics. Although these methods are proven successful for many applications, there is a grand challenge of improving their generalizability. This is particularly essential when data-driven models are employed within outer-loop applications like optimization. In this work, we put forth a physics-guided machine learning (PGML) framework that leverages the interpretable physics-based model with a deep learning model. Leveraging a concatenated neural network design from multi-modal data sources, the PGML framework is capable of enhancing the generalizability of data-driven models and effectively protects against or inform about the inaccurate predictions resulting from extrapolation. We apply the PGML framework as a novel model fusion approach combining the physics-based Galerkin projection model and long- to short-term memory (LSTM) network for parametric model order reduction of fluid flows. We demonstrate the improved generalizability of the PGML framework against a purely data-driven approach through the injection of physics features into intermediate LSTM layers. Our quantitative analysis shows that the overall model uncertainty can be reduced through the PGMLmore » approach, especially for test data coming from a distribution different than the training data. Moreover, we demonstrate that our approach can be used as an inverse diagnostic tool providing a confidence score associated with models and observations. The proposed framework also allows for multi-fidelity computing by making use of low-fidelity models in the online deployment of quantified data-driven models.« less

Authors:

Pawar, Suraj  ^[1];

• Search DOE PAGES for author "Pawar, Suraj"
• Search DOE PAGES for ORCID "0000-0001-7562-799X"

San, Omer  ^[1];

• Search DOE PAGES for author "San, Omer"
• Search DOE PAGES for ORCID "0000-0002-2241-4648"

Nair, Aditya ^[2]; Rasheed, Adil  ^[3];

• Search DOE PAGES for author "Rasheed, Adil"
• Search DOE PAGES for ORCID "0000-0003-2690-983X"

Kvamsdal, Trond  ^[3]

• Search DOE PAGES for author "Kvamsdal, Trond"
• Search DOE PAGES for ORCID "0000-0002-1615-7649"

---

+ Show Author Affiliations

1. Oklahoma State University, Stillwater, OK (United States)
2. University of Nevada, Reno, NV (United States)
3. Norwegian University of Science and Technology, Trondheim (Norway); SINTEF Energy Research, Trondheim (Norway)

Publication Date:

Tue Jun 29 00:00:00 EDT 2021

Research Org.:

Oklahoma State Univ., Stillwater, OK (United States)

Sponsoring Org.:

USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)

OSTI Identifier:

1853024

Alternate Identifier(s):

OSTI ID: 1804133

Grant/Contract Number:  

SC0019290

Resource Type:

Accepted Manuscript

Journal Name:

Physics of Fluids

Additional Journal Information:

Journal Volume: 33; Journal Issue: 6; Journal ID: ISSN 1070-6631

Publisher:

American Institute of Physics (AIP)

Country of Publication:

United States

Language:

English

Subject:

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSICS; 42 ENGINEERING; fluid mechanics; computational physics; equations of fluid dynamics; artificial neural networks; vortex dynamics; computational methods; machine learning; Navier Stokes equations