pith. sign in

arxiv: 2502.03672 · v2 · pith:CS75W5HKnew · submitted 2025-02-05 · ⚛️ physics.comp-ph · cs.LG· cs.NA· math.NA

Physically consistent predictive reduced-order modeling by enhancing Operator Inference with state constraints

Hyeonghun Kim , Boris Kramer This is my paper

Pith reviewed 2026-05-23 03:32 UTC · model grok-4.3

classification ⚛️ physics.comp-ph cs.LGcs.NAmath.NA
keywords reduced-order modelingOperator Inferencestate constraintschar combustionscientific machine learningphysical consistencyextrapolationstability
0
0 comments X

The pith

Embedding state constraints into Operator Inference yields stable reduced-order models that extrapolate over 200 percent past training data for char combustion.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper seeks to establish that augmenting Operator Inference with embedded physical state constraints during the learning process produces reduced-order models with improved stability and accuracy for high-dimensional nonlinear systems. This is paired with a new method for selecting regularization hyperparameters via a key performance indicator. A sympathetic reader would care because complex multiphysics simulations generate numerous state variables subject to physical limits, and reliable long-horizon predictions without prohibitive cost would support better analysis and design. The work demonstrates these gains specifically for char combustion, where the constrained models outperform standard Operator Inference and other stability methods while remaining computationally efficient.

Core claim

The paper claims that embedding state constraints in the Operator Inference reduced-order model predictions, together with performance-indicator-based hyperparameter selection, produces models superior to standard Operator Inference and other stability-enhancing approaches in stability and accuracy; for char combustion this enables extrapolation over 200 percent past the training regime while preserving computational efficiency and physical consistency.

What carries the argument

State constraints-embedded Operator Inference, which incorporates physical constraints on state variables directly into the learning of the low-dimensional representation from data.

If this is right

  • State predictions remain more stable and accurate than those from standard Operator Inference or other stability methods.
  • The models extrapolate over 200 percent past the training regime for the char combustion case.
  • Predictions stay physically consistent with the underlying system constraints.
  • The overall procedure remains computationally efficient compared with full-order simulation.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same constraint-embedding step could be tested on other multiphysics problems whose state variables obey known bounds to check whether stability gains transfer.
  • The key-performance-indicator approach to regularization might reduce trial-and-error tuning when Operator Inference is applied to new systems.
  • If the physical constraints are only approximately known, the method could still improve consistency but would require separate checks on bias introduced by inexact constraints.

Load-bearing premise

Embedding the state constraints during learning does not introduce new bias or instability that offsets the reported extrapolation gains, and the key performance indicator selects hyperparameters that generalize beyond the specific combustion dataset.

What would settle it

A direct comparison on a second multiphysics dataset or a longer extrapolation interval showing whether the state-constrained models lose physical consistency or accuracy relative to the unconstrained baselines.

Figures

Figures reproduced from arXiv: 2502.03672 by Boris Kramer, Hyeonghun Kim.

Figure 1
Figure 1. Figure 1: Computational domain of a laboratory-scale fluidized bed for char combustion. [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Convergence study. Mesh sizes nx ∈ {2520, 10800, 22400, 42000, 88200} are evaluated. The left y-axis corresponds to integrated mass fractions for O2 and CO2. The right y-axis corresponds to CPU hours on a logarithmic scale. 3. Operator Inference reduced-order model with state constraints The char combustion model is governed by the gas phase conservation equations (1)–(4), the solid phase conservation equa… view at source ↗
Figure 3
Figure 3. Figure 3: Schematic of the state propagation in the ROM with species limiters, which are state [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Interpretation of the state constraints-embedded state propagation, as presented in Figure [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Interpretation of state-constraints-embedded state propagation, when using a block [PITH_FULL_IMAGE:figures/full_fig_p016_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Projection errors of each variable. The errors are computed with the variables in their [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Comparison of two ROM predictions of temperature and N [PITH_FULL_IMAGE:figures/full_fig_p018_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of spatially averaged pointwise error [PITH_FULL_IMAGE:figures/full_fig_p021_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The sum of mass fractions over time. The sum is computed by summing five mass fractions [PITH_FULL_IMAGE:figures/full_fig_p023_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Comparison of thermal energy rate and thermal energy. The shaded area under the [PITH_FULL_IMAGE:figures/full_fig_p023_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Comparison across four training horizons: [PITH_FULL_IMAGE:figures/full_fig_p025_11.png] view at source ↗
read the original abstract

Numerical simulations of complex multiphysics systems, such as char combustion considered herein, yield numerous state variables that inherently exhibit physical constraints. This paper presents a new approach to augment Operator Inference -- a methodology within scientific machine learning that enables learning from data a low-dimensional representation of a high-dimensional system governed by nonlinear partial differential equations -- by embedding such state constraints in the reduced-order model predictions. In the model learning process, we propose a new way to choose regularization hyperparameters based on a key performance indicator. Since embedding state constraints improves the stability of the Operator Inference reduced-order model, we compare the proposed state constraints-embedded Operator Inference with the standard Operator Inference and other stability-enhancing approaches. For an application to char combustion, we demonstrate that the proposed approach yields state predictions superior to the other methods regarding stability and accuracy. It extrapolates over 200\% past the training regime while being computationally efficient and physically consistent.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 0 minor

Summary. The paper augments Operator Inference (OpInf) for reduced-order modeling of nonlinear PDE-governed systems by embedding physical state constraints during the learning stage and selecting regularization hyperparameters via a key performance indicator (KPI). On a char-combustion application, the resulting constrained OpInf model is reported to outperform standard OpInf and other stability-enhancing variants in stability and accuracy, to extrapolate stably more than 200% beyond the training regime, and to remain physically consistent while remaining computationally efficient.

Significance. If the central claims are substantiated, the work would supply a practical route to enforce physical consistency inside data-driven reduced-order models without sacrificing the non-intrusive character of OpInf. The reported 200% extrapolation on a multiphysics combustion problem would constitute a concrete, falsifiable demonstration of improved predictive capability for systems whose states obey hard constraints.

major comments (3)
  1. [Abstract] Abstract and results section: the claim that constraint embedding 'improves the stability of the Operator Inference reduced-order model' is presented without a derivation or numerical verification that the added constraints preserve the underlying least-squares learning guarantees of OpInf; no error bars, cross-validation statistics, or sensitivity analysis on the constraint embedding are supplied.
  2. [Abstract] Abstract and hyperparameter-selection paragraph: the KPI used to choose regularization hyperparameters is not shown to be independent of the test trajectories; without an ablation that compares KPI selection against, e.g., cross-validation on held-out training windows, the reported 200% extrapolation gain cannot be distinguished from dataset-specific tuning.
  3. [Abstract] Comparison paragraph: the statement that the proposed method 'yields state predictions superior to the other methods regarding stability and accuracy' is not accompanied by quantitative tables or figures that report the precise metrics, the number of independent runs, or the precise definition of the 200% extrapolation horizon.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments. We agree that the abstract and results require additional substantiation and will revise the manuscript to include a theoretical note on the least-squares properties, an ablation study for the KPI, and quantitative tables with precise metrics and definitions. Point-by-point responses are below.

read point-by-point responses

  1. Referee: [Abstract] Abstract and results section: the claim that constraint embedding 'improves the stability of the Operator Inference reduced-order model' is presented without a derivation or numerical verification that the added constraints preserve the underlying least-squares learning guarantees of OpInf; no error bars, cross-validation statistics, or sensitivity analysis on the constraint embedding are supplied.

    Authors: Stability improvement is demonstrated numerically on the char combustion problem, where constrained models remain bounded while others diverge. We acknowledge that a formal derivation preserving least-squares guarantees is absent; the constraints are added as linear equalities to the OpInf regression, preserving convexity. We will add a methods discussion on this point plus error bars from repeated runs, cross-validation statistics, and sensitivity analysis on constraint weights. revision: yes

  2. Referee: [Abstract] Abstract and hyperparameter-selection paragraph: the KPI used to choose regularization hyperparameters is not shown to be independent of the test trajectories; without an ablation that compares KPI selection against, e.g., cross-validation on held-out training windows, the reported 200% extrapolation gain cannot be distinguished from dataset-specific tuning.

    Authors: The KPI is evaluated exclusively on training trajectories using physical consistency indicators, ensuring independence from test data. To strengthen the claim, we will add an ablation comparing KPI selection against cross-validation on held-out training windows in the revised manuscript. revision: yes

  3. Referee: [Abstract] Comparison paragraph: the statement that the proposed method 'yields state predictions superior to the other methods regarding stability and accuracy' is not accompanied by quantitative tables or figures that report the precise metrics, the number of independent runs, or the precise definition of the 200% extrapolation horizon.

    Authors: Figures illustrate the trajectories, but we agree precise metrics are needed. The revision will add a results table with stability/accuracy metrics (e.g., L2 errors), the number of independent runs, and an explicit definition of the 200% horizon as simulation times reaching three times the training duration. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation or claims

full rationale

The paper augments Operator Inference by embedding state constraints during learning and selects regularization hyperparameters via a proposed key performance indicator. These steps are presented as methodological choices whose outputs (reduced-order model predictions) are then evaluated empirically on char combustion data for stability, accuracy, and extrapolation. No equation or step is shown to reduce the reported performance metrics to the inputs by construction, nor does any load-bearing claim rely on a self-citation chain, uniqueness theorem imported from the authors, or an ansatz smuggled via prior work. The extrapolation results are treated as independent validation rather than a fitted quantity renamed as prediction. The derivation chain therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Review performed on abstract only; no explicit free parameters, axioms, or invented entities are stated. The method implicitly assumes that state constraints can be expressed as linear or simple inequalities compatible with the reduced basis and that the KPI is a reliable proxy for out-of-sample stability.

pith-pipeline@v0.9.0 · 5688 in / 1046 out tokens · 22245 ms · 2026-05-23T03:32:30.749003+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

54 extracted references · 54 canonical work pages

  1. [1]

    Adibnazari, H

    I. Adibnazari, H. Sharma, J. C. Torralba, B. Kramer, and M. T. Tolley. Full-body optimal control of a swimming soft robot enabled by data-driven model reduction.2023 Southern California Robotics (SCR) Symposium, 2023

    work page 2023

  2. [2]

    Agrawal, R

    A. Agrawal, R. Verschueren, S. Diamond, and S. Boyd. A rewriting system for convex optimiza- tion problems.Journal of Control and Decision, 5(1):42–60, 2018

    work page 2018

  3. [3]

    S. E. Ahmed, S. Pawar, O. San, A. Rasheed, T. Iliescu, and B. R. Noack. On closures for reduced order models—a spectrum of first-principle to machine-learned avenues.Physics of Fluids, 33(9), 2021

    work page 2021

  4. [4]

    Basu.Combustion and Gasification in Fluidized Beds

    P. Basu.Combustion and Gasification in Fluidized Beds. CRC press, 2006

    work page 2006

  5. [5]

    Baumgart and G

    A. Baumgart and G. Blanquart. Ensuring P s ys = 1in transport of species mass fractions. Journal of Computational Physics, 513:113199, 2024

    work page 2024

  6. [6]

    Benner, P

    P. Benner, P. Goyal, J. Heiland, and I. P. Duff. Operator inference and physics-informed learn- ing of low-dimensional models for incompressible flows.Electronic Transactions on Numerical Analysis, 56:28–51, 2022

    work page 2022

  7. [7]

    Benner, P

    P. Benner, P. Goyal, B. Kramer, B. Peherstorfer, and K. Willcox. Operator inference for non- intrusive model reduction of systems with non-polynomial nonlinear terms.Computer Methods in Applied Mechanics and Engineering, 372:113433, 2020

    work page 2020

  8. [8]

    Benner, S

    P. Benner, S. Gugercin, and K. Willcox. A survey of projection-based model reduction methods for parametric dynamical systems.SIAM Review, 57(4):483–531, 2015

    work page 2015

  9. [9]

    Benosman, J

    M. Benosman, J. Borggaard, and B. Kramer. Robust POD model stabilization for the 3D Boussi- nesq equations based on Lyapunov theory and extremum seeking. InAmerican Control Confer- ence (ACC), pages 1827–1832. IEEE, 2017

    work page 2017

  10. [10]

    Benosman, J

    M. Benosman, J. Borggaard, O. San, and B. Kramer. Learning-based robust stabilization for reduced-order models of 2D and 3D Boussinesq equations.Applied Mathematical Modelling, 49:162–181, 2017

    work page 2017

  11. [11]

    M. A. Clarke and J. M. Musser. The MFiX particle-in-cell method (MFiX-PIC) theory guide. Technical report, National Energy Technology Laboratory (NETL), 2020

    work page 2020

  12. [12]

    Diamond and S

    S. Diamond and S. Boyd. CVXPY: A python-embedded modeling language for convex optimiza- tion.The Journal of Machine Learning Research, 17(1):2909–2913, 2016

    work page 2016

  13. [13]

    Dryer and I

    F. Dryer and I. Glassman. High-temperature oxidation of CO and CH4.Symposium (Interna- tional) on Combustion, 14(1):987–1003, 1973

    work page 1973

  14. [14]

    Farcas, R

    I.-G. Farcas, R. Gundevia, R. Munipalli, and K. E. Willcox. Parametric non-intrusive reduced- order models via operator inference for large-scale rotating detonation engine simulations. In AIAA SciTech 2023 Forum, page 0172, 2023

    work page 2023

  15. [15]

    Filanova, I

    Y. Filanova, I. P. Duff, P. Goyal, and P. Benner. An operator inference oriented approach for linear mechanical systems.Mechanical Systems and Signal Processing, 200:110620, 2023

    work page 2023

  16. [16]

    Geng and D

    Y. Geng and D. Che. An extended DEM–CFD model for char combustion in a bubbling fluidized bed combustor of inert sand.Chemical Engineering Science, 66(2):207–219, 2011. 30

    work page 2011

  17. [17]

    Y. Geng, J. Singh, L. Ju, B. Kramer, and Z. Wang. Gradient preserving operator inference: Data-driven reduced-order models for equations with gradient structure.Computer Methods in Applied Mechanics and Engineering, 427:117033, 2024

    work page 2024

  18. [18]

    Ghattas and K

    O. Ghattas and K. Willcox. Learning physics-based models from data: perspectives from inverse problems and model reduction.Acta Numerica, 30:445–554, 2021

    work page 2021

  19. [19]

    Gruber and I

    A. Gruber and I. Tezaur. Canonical and noncanonical Hamiltonian operator inference.Computer Methods in Applied Mechanics and Engineering, 416:116334, 2023

    work page 2023

  20. [20]

    Halko, P.-G

    N. Halko, P.-G. Martinsson, and J. A. Tropp. Finding structure with randomness: Probabilistic algorithms for constructing approximate matrix decompositions.SIAM Review, 53(2):217–288, 2011

    work page 2011

  21. [21]

    J. S. Hesthaven, G. Rozza, B. Stamm, et al.Certified Reduced Basis Methods for Parametrized Partial Differential Equations, volume 590. Springer, 2016

    work page 2016

  22. [22]

    Huang, K

    C. Huang, K. Duraisamy, and C. Merkle. Data-informed species limiters for local robustness control of reduced-order models of reacting flow. InAIAA SciTech 2020 Forum, 2020

    work page 2020

  23. [23]

    Issan and B

    O. Issan and B. Kramer. Predicting solar wind streams from the inner-heliosphere to earth via shifted operator inference.Journal of Computational Physics, 473:111689, 2023

    work page 2023

  24. [24]

    P. Jain, S. McQuarrie, and B. Kramer. Performance comparison of data-driven reduced models for a single-injector combustion process. InAIAA Propulsion and Energy 2021 Forum, page 3633, 2021

    work page 2021

  25. [25]

    Kalashnikova, B

    I. Kalashnikova, B. van Bloemen Waanders, S. Arunajatesan, and M. Barone. Stabilization of projection-based reduced order models for linear time-invariant systems via optimization-based eigenvalue reassignment.Computer Methods in Applied Mechanics and Engineering, 272:251–270, 2014

    work page 2014

  26. [26]

    Anintrinsicstabilizationschemeforproperorthogonaldecomposition based low-dimensional models.Physics of Fluids, 19(5), 2007

    V.L.KalbandA.E.Deane. Anintrinsicstabilizationschemeforproperorthogonaldecomposition based low-dimensional models.Physics of Fluids, 19(5), 2007

    work page 2007

  27. [27]

    A. A. Kaptanoglu, J. L. Callaham, A. Aravkin, C. J. Hansen, and S. L. Brunton. Promoting global stability in data-driven models of quadratic nonlinear dynamics.Physical Review Fluids, 6(9):094401, 2021

    work page 2021

  28. [28]

    Koike and E

    T. Koike and E. Qian. Energy-preserving reduced operator inference for efficient design and control. InAIAA SciTech 2024 Forum, page 1012, 2024

    work page 2024

  29. [29]

    B. Kramer. Stability domains for quadratic-bilinear reduced-order models.SIAM Journal on Applied Dynamical Systems, 20(2):981–996, 2021

    work page 2021

  30. [30]

    Kramer, B

    B. Kramer, B. Peherstorfer, and K. E. Willcox. Learning nonlinear reduced models from data with operator inference.Annual Review of Fluid Mechanics, 56(1):521–548, 2024

    work page 2024

  31. [31]

    X. Ku, T. Li, and T. Løvås. CFD–DEM simulation of biomass gasification with steam in a fluidized bed reactor.Chemical Engineering Science, 122:270–283, 2015

    work page 2015

  32. [32]

    La Nauze, K

    R. La Nauze, K. Jung, and J. Kastl. Mass transfer to large particles in fluidised beds of smaller particles.Chemical Engineering Science, 39(11):1623–1633, 1984

    work page 1984

  33. [33]

    D. Lee, E. Lavichant, and B. Kramer. Global sensitivity analysis with limited data via sparsity- promoting D-MORPH regression: Application to char combustion.Journal of Computational Physics, 511:113116, 2024. 31

    work page 2024

  34. [34]

    S. A. McQuarrie, C. Huang, and K. E. Willcox. Data-driven reduced-order models via regularised operator inference for a single-injector combustion process.Journal of the Royal Society of New Zealand, 51(2):194–211, 2021

    work page 2021

  35. [35]

    S. A. McQuarrie, P. Khodabakhshi, and K. E. Willcox. Nonintrusive reduced-order models for parametric partial differential equations via data-driven operator inference.SIAM Journal on Scientific Computing, 45(4):A1917–A1946, 2023

    work page 2023

  36. [36]

    J. M. Musser and J. E. Carney. Theoretical review of the MFiX fluid and two-fluid models. Technical report, National Energy Technology Laboratory (NETL), 2020

    work page 2020

  37. [37]

    Pedregosa, G

    F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, A. Müller, J. Nothman, G. Louppe, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and Édouard Duchesnay. Scikit-learn: Ma- chine learning in python, 2018

    work page 2018

  38. [38]

    Peherstorfer and K

    B. Peherstorfer and K. Willcox. Data-driven operator inference for nonintrusive projection-based model reduction.Computer Methods in Applied Mechanics and Engineering, 306:196–215, 2016

    work page 2016

  39. [39]

    Qian, I.-G

    E. Qian, I.-G. Farcas, and K. Willcox. Reduced operator inference for nonlinear partial differential equations.SIAM Journal on Scientific Computing, 44(4):A1934–A1959, 2022

    work page 2022

  40. [40]

    Quarteroni, G

    A. Quarteroni, G. Rozza, et al.Reduced Order Methods for Modeling and Computational Reduc- tion, volume 9. Springer, 2014

    work page 2014

  41. [41]

    P. R. B. Rocha, J. L. de Sousa Almeida, M. S. de Paula Gomes, and A. C. N. Junior. Reduced- order modeling of the two-dimensional Rayleigh–Bénard convection flow through a non-intrusive operator inference.Engineering Applications of Artificial Intelligence, 126:106923, 2023

    work page 2023

  42. [42]

    Sawant, B

    N. Sawant, B. Kramer, and B. Peherstorfer. Physics-informed regularization and structure preser- vation for learning stable reduced models from data with operator inference.Computer Methods in Applied Mechanics and Engineering, 404:115836, 2023

    work page 2023

  43. [43]

    F. Scala. Mass transfer around freely moving active particles in the dense phase of a gas fluidized bed of inert particles.Chemical Engineering Science, 62(16):4159–4176, 2007

    work page 2007

  44. [44]

    Schlegel and B

    M. Schlegel and B. R. Noack. On long-term boundedness of Galerkin models.Journal of Fluid Mechanics, 765:325–352, 2015

    work page 2015

  45. [45]

    Sharma and B

    H. Sharma and B. Kramer. Preserving Lagrangian structure in data-driven reduced-order mod- eling of large-scale dynamical systems.Physica D: Nonlinear Phenomena, 462:134128, 2024

    work page 2024

  46. [46]

    Sharma, D

    H. Sharma, D. A. Najera-Flores, M. D. Todd, and B. Kramer. Lagrangian operator inference enhanced with structure-preserving machine learning for nonintrusive model reduction of me- chanical systems.Computer Methods in Applied Mechanics and Engineering, 423:116865, 2024

    work page 2024

  47. [47]

    Sharma, Z

    H. Sharma, Z. Wang, and B. Kramer. Hamiltonian operator inference: Physics-preserving learn- ing of reduced-order models for canonical Hamiltonian systems.Physica D: Nonlinear Phenom- ena, 431:133122, 2022

    work page 2022

  48. [48]

    Swischuk, B

    R. Swischuk, B. Kramer, C. Huang, and K. Willcox. Learning physics-based reduced-order models for a single-injector combustion process.AIAA Journal, 58(6):2658–2672, June 2020

    work page 2020

  49. [49]

    Z. Wang, I. Akhtar, J. Borggaard, and T. Iliescu. Proper orthogonal decomposition closure models for turbulent flows: A numerical comparison.Computer Methods in Applied Mechanics and Engineering, 237:10–26, 2012. 32

    work page 2012

  50. [50]

    Winter, M

    F. Winter, M. E. Prah, and H. Hofbauer. Temperatures in a fuel particle burning in a flu- idized bed: The effect of drying, devolatilization, and char combustion.Combustion and Flame, 108(3):302–314, 1997

    work page 1997

  51. [51]

    J. Xie, W. Zhong, and Y. Shao. Study on the char combustion in a fluidized bed by CFD-DEM simulations: Influences of fuel properties.Powder Technology, 394:20–34, 2021

    work page 2021

  52. [52]

    S. Yang, H. Wang, Y. Wei, J. Hu, and J. W. Chew. Particle-scale modeling of biomass gasification in the three-dimensional bubbling fluidized bed.Energy Conversion and Management, 196:1–17, 2019

    work page 2019

  53. [53]

    Yıldız, P

    S. Yıldız, P. Goyal, P. Benner, and B. Karasözen. Learning reduced-order dynamics for parametrized shallow water equations from data.International Journal for Numerical Methods in Fluids, 93(8):2803–2821, 2021

    work page 2021

  54. [54]

    B. G. Zastrow, A. Chaudhuri, K. Willcox, A. S. Ashley, and M. C. Henson. Data-driven model reduction via operator inference for coupled aeroelastic flutter. InAIAA SciTech 2023 Forum, page 0330, 2023. 33

    work page 2023