pith. sign in

arxiv: 2605.20501 · v1 · pith:JNIZSE57new · submitted 2026-05-19 · ⚛️ physics.comp-ph · cs.NA· math.NA· math.OC

Adaptive Multi-Fidelity Structural Optimization under Fluid-Structure Interaction

Aditya Narkhede , Erick Rivas , Kevin Wang This is my paper

Pith reviewed 2026-05-21 05:56 UTC · model grok-4.3

classification ⚛️ physics.comp-ph cs.NAmath.NAmath.OC
keywords fluid-structure interactionmulti-fidelity optimizationsurrogate modelingshape optimizationGaussian processstructural optimizationshock loading
0
0 comments X

The pith

An adaptive surrogate for fluid loads reduces structural optimization cost by 80 percent while keeping final designs within 2.3 percent of full high-fidelity accuracy.

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

The paper develops an optimization procedure for flexible structures under fluid loads that avoids running the expensive coupled fluid solver for most candidate designs. It maintains an on-the-fly surrogate of fluid-induced pressures built from nearest-neighbor search and radial interpolation of completed simulations, then uses an adaptive Gaussian process to estimate how large the surrogate error is likely to be. When the estimated error is low, the optimizer proceeds with the cheap surrogate; otherwise it calls the full high-fidelity fluid analysis. As the search converges, the growing cluster of nearby designs improves the surrogate automatically, so fewer full simulations are needed. The final best design is always checked with the complete high-fidelity model. Readers would care because repeated fluid-structure simulations dominate the cost of many vehicle and structural design loops, and this approach removes the need for separate offline training.

Core claim

The central claim is that an adaptive multi-fidelity framework, which incrementally updates a non-intrusive surrogate for fluid-induced loads via nearest-neighbor search and radial interpolation and employs an adaptive Gaussian process to predict surrogate error, permits risk-aware selection between surrogate and full coupled FSI evaluations. This yields an 80 percent reduction in computational cost for shape optimization of a flexible panel under shock loading while the final design remains within 2.3 percent of the result from fully high-fidelity optimization. The approach is supported by a theoretical argument on a simplified model problem showing that the leading-order error is a monoton

What carries the argument

The adaptive surrogate for fluid loads based on nearest-neighbor search and radial interpolation, updated incrementally, together with an adaptive Gaussian process regression model that predicts surrogate error to enable safe skipping of high-fidelity fluid evaluations.

If this is right

  • The surrogate accuracy improves automatically as design points cluster near the optimum, reducing the fraction of full fluid solves needed.
  • No separate offline training phase is required.
  • The underlying high-fidelity structural model is retained for every design evaluation.
  • The final reported design is always obtained from a complete high-fidelity FSI analysis.
  • A simplified model problem shows that the leading-order error grows monotonically yet remains bounded with increasing fluid added mass.

Where Pith is reading between the lines

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

  • The same incremental surrogate-plus-uncertainty approach could be applied to other expensive physics couplings, such as aeroelastic optimization of wings, where one solver dominates runtime.
  • If the Gaussian process error predictions remain reliable on problems with more design variables, the fraction of skipped high-fidelity calls could rise further.
  • The hybrid Lagrangian-Eulerian load transfer and local basis decomposition for surface orientation may need additional checks when shape changes become very large.

Load-bearing premise

The surrogate model and its Gaussian-process uncertainty estimates remain accurate enough throughout the optimization to allow many high-fidelity fluid evaluations to be skipped without producing a noticeably inferior final design.

What would settle it

Run the same flexible-panel shape optimization once with the full adaptive method and once with high-fidelity fluid analysis required at every step; if the reported 80 percent cost reduction disappears or the final objective value differs by more than 2.3 percent, the central claim is falsified.

Figures

Figures reproduced from arXiv: 2605.20501 by Aditya Narkhede, Erick Rivas, Kevin Wang.

Figure 1
Figure 1. Figure 1: Fundamental concepts in FSI-constrained structural optimization: independent variables (space [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A two-dimensional fluid-structure interaction model problem. [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustration of spatial interpolation across di [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Illustration of the local orthonormal basis used for traction decomposition across structural designs. The normal traction component [PITH_FULL_IMAGE:figures/full_fig_p015_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Implementation of the baseline optimization framework (SOFICS-Base). [PITH_FULL_IMAGE:figures/full_fig_p019_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Implementation of the adaptive multi-fidelity optimization framework (SOFICS-AMF). [PITH_FULL_IMAGE:figures/full_fig_p023_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Piston in a one-dimensional flow. (a) Problem setup. (b) Fluid pressure field for [PITH_FULL_IMAGE:figures/full_fig_p025_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Evolution of the design populations in BASE. (a) Broader view of the design space; (b) detailed view in the vicinity of the optimum. The final design is marked with a pentagram. For reference, the two linear constraints are also shown, and the true optimum lies at their intersection [PITH_FULL_IMAGE:figures/full_fig_p027_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Evolution of the design populations in AMF. (a) Broader view of the design space; (b) detailed view in the vicinity of the optimum. Designs evaluated using the surrogate model are indicated by black outlines, and the final design is marked with a pentagram [PITH_FULL_IMAGE:figures/full_fig_p027_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: A flexible cantilever panel subject to shock loading. [PITH_FULL_IMAGE:figures/full_fig_p028_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Comparison of current simulation result with published data for a reference design ( [PITH_FULL_IMAGE:figures/full_fig_p029_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Comparison of the three methods BASE, AMF, and DRY. (a) Iteration history of the merit function; (b) evolution of the optimal design at four iterations. The error in α is calculated using the final result from BASE as reference [PITH_FULL_IMAGE:figures/full_fig_p032_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Design population at selected iterations of the [PITH_FULL_IMAGE:figures/full_fig_p033_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Pressure time histories recorded at the panel’s tip across selected iterations of the [PITH_FULL_IMAGE:figures/full_fig_p033_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Evolution of the design population in the [PITH_FULL_IMAGE:figures/full_fig_p034_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Comparison of the dynamic responses of the optimal designs obtained from [PITH_FULL_IMAGE:figures/full_fig_p035_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Surrogate error and uncertainty at 21 points in the design space as estimated by the GPR model, based on CFD-CSD simulation data [PITH_FULL_IMAGE:figures/full_fig_p036_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Cost comparison for the three studies BASE, DRY, and AMF. The number of design evaluations is shown on the left axis; the total computation time is shown on the right axis. 6. Concluding remarks New high-fidelity analysis capabilities do not directly translate into new design optimization capabilities, as their computational cost is often too high for many-query applications. This is presently the case fo… view at source ↗
read the original abstract

The design of structures and vehicles subject to fluid-structure interaction (FSI) often requires high-fidelity coupled analysis. While the design variables pertain to the structure, the computational cost is dominated by the fluid solver, making iterative optimization prohibitively expensive. This paper presents an adaptive multi-fidelity optimization method combining high-fidelity FSI analysis with a lightweight surrogate for fluid-induced loads and a decision model that selects between surrogate and high-fidelity fluid evaluations. During optimization, completed FSI analyses incrementally update a non-intrusive surrogate model based on nearest-neighbor search and radial interpolation. A hybrid Lagrangian-Eulerian mapping function is developed to transfer fluid loads between structural designs. The evolution of surface orientation is handled by decomposing the traction vectors into local orthonormal bases. An adaptive Gaussian process regression model is employed to predict surrogate error and quantify uncertainty, allowing risk-aware selection of when coupled analysis is required. As design evaluations cluster near the optimum, the accuracy of the surrogate model naturally improves, thereby reducing the reliance on the fluid solver. It requires no offline training, preserves the high-fidelity structural model in all design evaluations, and ensures that the final design is evaluated by high-fidelity FSI analysis. The fundamental idea is justified theoretically using a simplified model problem, which shows that the leading-order error is a monotonically increasing, concave, and bounded function of the fluid added mass. The framework is demonstrated on two benchmark problems. For shape optimization of a flexible panel under shock loading, results show an $80\%$ reduction in computational cost while maintaining accuracy within $2.3\%$ of fully high-fidelity FSI optimization.

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

2 major / 2 minor

Summary. The paper proposes an adaptive multi-fidelity optimization framework for structural design under fluid-structure interaction (FSI). It combines high-fidelity coupled FSI analysis with an incrementally updated non-intrusive surrogate for fluid-induced loads (nearest-neighbor search plus radial interpolation), a hybrid Lagrangian-Eulerian load-transfer mapping that decomposes tractions into local orthonormal bases, and an adaptive Gaussian process regression model that predicts surrogate error to enable risk-aware selection between surrogate and high-fidelity fluid evaluations. A theoretical argument on a simplified model shows that the leading-order error is monotonically increasing, concave, and bounded in fluid added mass. The method requires no offline training, preserves the high-fidelity structural model at every design point, and guarantees that the final optimum is evaluated with full high-fidelity FSI. Benchmark results on shape optimization of a flexible panel under shock loading report an 80% reduction in computational cost while remaining within 2.3% of the fully high-fidelity optimum.

Significance. If the benchmark claims hold under detailed scrutiny, the approach could meaningfully lower the barrier to performing structural optimization in FSI-dominated regimes (e.g., aerospace panels, flexible aircraft). Strengths include the incremental surrogate update that improves automatically as designs cluster near the optimum, the absence of any offline training requirement, and the explicit guarantee that the reported optimum is always high-fidelity. The simplified-model error analysis provides a useful, if limited, theoretical anchor for the uncertainty-driven skipping strategy.

major comments (2)
  1. [§5] §5 (flexible-panel benchmark results): the stated 80% cost reduction and 2.3% accuracy figures are presented without error bars, iteration histories, or tabulated comparisons against the pure high-fidelity baseline; these omissions prevent assessment of whether the adaptive GP uncertainty model reliably prevented premature skipping of fluid solves during early iterations when geometry changes are largest.
  2. [§3.2–3.3] §3.2–3.3 (surrogate construction and decision rule): the nearest-neighbor + radial-interpolation surrogate together with the adaptive GP error predictor is described at a high level, but no explicit threshold or risk metric is given for deciding when to invoke the high-fidelity fluid solver; without this, it is impossible to verify that the “safe skipping” claim is robust to the monotonicity and boundedness properties asserted in the simplified-model analysis.
minor comments (2)
  1. [§3.1] The hybrid Lagrangian-Eulerian mapping is introduced without an accompanying equation or pseudocode block; adding a compact algorithmic description would improve reproducibility.
  2. [Figures 4–6] Figure captions for the optimization convergence plots should explicitly state the number of high-fidelity versus surrogate evaluations performed in each run.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major comment below and will revise the manuscript accordingly to improve transparency and verifiability of the results.

read point-by-point responses

  1. Referee: [§5] §5 (flexible-panel benchmark results): the stated 80% cost reduction and 2.3% accuracy figures are presented without error bars, iteration histories, or tabulated comparisons against the pure high-fidelity baseline; these omissions prevent assessment of whether the adaptive GP uncertainty model reliably prevented premature skipping of fluid solves during early iterations when geometry changes are largest.

    Authors: We agree that additional detail is needed to allow assessment of the adaptive strategy. In the revised manuscript we will add a table of per-iteration high-fidelity versus surrogate counts, convergence histories for both the adaptive and pure high-fidelity runs, and error bars (or sensitivity ranges) on the reported cost and accuracy metrics. These additions will show that the GP uncertainty model avoided premature skipping while geometry changes remained large. revision: yes

  2. Referee: [§3.2–3.3] §3.2–3.3 (surrogate construction and decision rule): the nearest-neighbor + radial-interpolation surrogate together with the adaptive GP error predictor is described at a high level, but no explicit threshold or risk metric is given for deciding when to invoke the high-fidelity fluid solver; without this, it is impossible to verify that the “safe skipping” claim is robust to the monotonicity and boundedness properties asserted in the simplified-model analysis.

    Authors: We acknowledge that the precise risk metric and threshold are not stated explicitly. The revised sections 3.2–3.3 will define the decision rule as skipping the high-fidelity fluid solve when the GP predictive standard deviation falls below a fixed fraction of the current objective value (the exact fraction used in the experiments will be reported). A short paragraph will also connect this rule to the monotonicity and boundedness results from the simplified model to substantiate the robustness claim. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper presents an adaptive multi-fidelity optimization framework whose central elements (incremental nearest-neighbor/radial surrogate updates, hybrid load mapping, adaptive GP error model, and risk-aware switching) are constructed as independent algorithmic components. The theoretical justification rests on a separate simplified model problem whose error properties (monotonic, concave, bounded in added mass) are derived directly rather than fitted to the target optimization results. Validation occurs via explicit benchmark comparisons that report concrete cost and accuracy deltas against fully high-fidelity runs, with the final design always evaluated at high fidelity. No load-bearing step reduces by the paper's own equations or self-citation to a tautological redefinition or pre-fitted parameter presented as a prediction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Based solely on the abstract, no explicit free parameters, axioms, or invented entities are detailed; the approach builds on standard techniques such as Gaussian process regression and nearest-neighbor interpolation without introducing new postulated entities or ad-hoc axioms beyond the described mapping function.

pith-pipeline@v0.9.0 · 5835 in / 1268 out tokens · 50696 ms · 2026-05-21T05:56:21.452048+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

54 extracted references · 54 canonical work pages

  1. [1]

    Motley, Z

    M. Motley, Z. Liu, Y . Young, Utilizing fluid–structure interactions to improve energy efficiency of composite marine propellers in spatially varying wake, Composite structures 90 (3) (2009) 304–313

    work page 2009

  2. [2]

    Kim, O.-K

    D.-H. Kim, O.-K. Lim, E.-H. Choi, Y . Noh, Optimization of 5-mw wind turbine blade using fluid structure interaction analysis, Journal of Mechanical Science and Technology 31 (2) (2017) 725–732

    work page 2017

  3. [3]

    R. Kano, K. Ryuzono, S. Date, Y . Abe, T. Okabe, Structural optimization of composite aircraft wing considering fluid–structure interaction and damage tolerance assessment using continuum damage mechanics, Aerospace Science and Technology (2025) 110652

    work page 2025

  4. [4]

    Chung, S

    H. Chung, S. Cao, M. Philen, P. S. Beran, K. G. Wang, Cfd-csd coupled analysis of underwater propulsion using a biomimetic fin-and-joint system, Computers & Fluids 172 (2018) 54–66

    work page 2018

  5. [5]

    A. L. Marsden, Optimization in cardiovascular modeling, Annual review of fluid mechanics 46 (1) (2014) 519– 546

    work page 2014

  6. [6]

    C. Long, A. Marsden, Y . Bazilevs, Shape optimization of pulsatile ventricular assist devices using fsi to minimize thrombotic risk, Computational Mechanics 54 (4) (2014) 921–932

    work page 2014

  7. [7]

    W. Ma, X. Zhao, C. Gilbert, K. Wang, Computational analysis of bubble–structure interactions in near-field underwater explosion, International Journal of Solids and Structures 242 (2022) 111527. 38

    work page 2022

  8. [8]

    Narkhede, S

    A. Narkhede, S. Islam, X. Sun, K. Wang, Fluid–structure coupled simulation framework for lightweight ex- plosion containment structures under large deformations, International Journal of Impact Engineering (2025) 105238doi:10.1016/j.ijimpeng.2025.105238

    work page doi:10.1016/j.ijimpeng.2025.105238 2025

  9. [9]

    Kamrin, C

    K. Kamrin, C. H. Rycroft, J.-C. Nave, Reference map technique for finite-strain elasticity and fluid–solid inter- action, Journal of the Mechanics and Physics of Solids 60 (11) (2012) 1952–1969

    work page 2012

  10. [10]

    Cerquaglia, D

    M.-L. Cerquaglia, D. Thomas, R. Boman, V . Terrapon, J.-P. Ponthot, A fully partitioned lagrangian framework for fsi problems characterized by free surfaces, large solid deformations and displacements, and strong added- mass effects, Computer Methods in Applied Mechanics and Engineering 348 (2019) 409–442

    work page 2019

  11. [11]

    K. Wang, A. Rallu, J.-F. Gerbeau, C. Farhat, Algorithms for interface treatment and load computation in embed- ded boundary methods for fluid and fluid–structure interaction problems, International Journal for Numerical Methods in Fluids 67 (9) (2011) 1175–1206. arXiv:https://onlinelibrary.wiley.com/doi/pdf/10.1002/fld.2556, doi:https://doi.org/10.1002/fl...

    work page doi:10.1002/fld.2556 2011

  12. [12]

    Kamensky, M.-C

    D. Kamensky, M.-C. Hsu, D. Schillinger, J. A. Evans, A. Aggarwal, Y . Bazilevs, M. S. Sacks, T. J. Hughes, An immersogeometric variational framework for fluid–structure interaction: Application to bioprosthetic heart valves, Computer methods in applied mechanics and engineering 284 (2015) 1005–1053

    work page 2015

  13. [13]

    S. Cao, A. Main, K. G. Wang, Robin-neumann transmission conditions for fluid-structure coupling: embedded boundary implementation and parameter analysis, International Journal for Numerical Methods in Engineering 115 (5) (2018) 578–603

    work page 2018

  14. [14]

    J. A. González, K. Park, Three-field partitioned analysis of fluid–structure interaction problems with a consistent interface model, Computer Methods in Applied Mechanics and Engineering 414 (2023) 116134

    work page 2023

  15. [15]

    B. A. Miller, J. J. McNamara, Efficient fluid-thermal-structural time marching with computational fluid dynam- ics, AIAA Journal 56 (9) (2018) 3610–3621

    work page 2018

  16. [16]

    Errera, F

    M.-P. Errera, F. Duchaine, Comparative study of coupling coefficients in dirichlet–robin procedure for fluid– structure aerothermal simulations, Journal of Computational Physics 312 (2016) 218–234

    work page 2016

  17. [17]

    Gravemeier, S

    V . Gravemeier, S. M. Civaner, W. A. Wall, A partitioned-monolithic finite element method for thermo-fluid– structure interaction, Computer Methods in Applied Mechanics and Engineering 401 (2022) 115596

    work page 2022

  18. [18]

    N. V . Queipo, R. T. Haftka, W. Shyy, T. Goel, R. Vaidyanathan, P. K. Tucker, Surrogate-based analysis and optimization, Progress in aerospace sciences 41 (1) (2005) 1–28. 39

    work page 2005

  19. [19]

    B. Liu, Q. Zhang, G. G. Gielen, A gaussian process surrogate model assisted evolutionary algorithm for medium scale expensive optimization problems, IEEE Transactions on Evolutionary Computation 18 (2) (2013) 180–192

    work page 2013

  20. [20]

    Zhang, S

    Y . Zhang, S. Hu, Y . Zhang, L. Chen, Optimization and analysis of centrifugal pump considering fluid-structure interaction, The scientific world journal 2014 (1) (2014) 131802

    work page 2014

  21. [21]

    Yoshimura, K

    M. Yoshimura, K. Shimoyama, T. Misaka, S. Obayashi, Topology optimization of fluid problems using genetic algorithm assisted by the kriging model, International Journal for Numerical Methods in Engineering 109 (4) (2017) 514–532

    work page 2017

  22. [22]

    M. C. Wu, D. Kamensky, C. Wang, A. J. Herrema, F. Xu, M. S. Pigazzini, A. Verma, A. L. Marsden, Y . Bazilevs, M.-C. Hsu, Optimizing fluid–structure interaction systems with immersogeometric analysis and surrogate mod- eling: Application to a hydraulic arresting gear, Computer Methods in Applied Mechanics and Engineering 316 (2017) 668–693

    work page 2017

  23. [23]

    A. F. Ribeiro, D. Casalino, C. Ferreira, Free wake panel method simulations of a highly flexible wing in flutter and gusts, Journal of Fluids and Structures 121 (2023) 103955

    work page 2023

  24. [24]

    Ortega, R

    E. Ortega, R. Flores, E. Cuartero, E. Oñate, Efficient aeroelastic analysis of inflatable structures using enhanced potential flow aerodynamics, Journal of Fluids and Structures 90 (2019) 230–245

    work page 2019

  25. [25]

    L. Ma, Y . Hu, J. Zheng, G. Deng, Y . Chen, Failure analysis for cylindrical explosion containment vessels, Engineering Failure Analysis 17 (5) (2010) 1221–1229

    work page 2010

  26. [26]

    M. J. Zahr, C. Farhat, Progressive construction of a parametric reduced-order model for pde-constrained opti- mization, International Journal for Numerical Methods in Engineering 102 (5) (2015) 1111–1135

    work page 2015

  27. [27]

    Tezaur, F

    R. Tezaur, F. As’ad, C. Farhat, Robust and globally efficient reduction of parametric, highly nonlinear computa- tional models and real time online performance, Computer Methods in Applied Mechanics and Engineering 399 (2022) 115392

    work page 2022

  28. [28]

    T. Lieu, C. Farhat, M. Lesoinne, Reduced-order fluid/structure modeling of a complete aircraft configuration, Computer Methods in Applied Mechanics and Engineering 195 (41) (2006) 5730–5742, john H. Argyris Memo- rial Issue. Part II. doi:https://doi.org/10.1016/j.cma.2005.08.026. URLhttps://www.sciencedirect.com/science/article/pii/S0045782505005153

    work page doi:10.1016/j.cma.2005.08.026 2006

  29. [29]

    D. Xiao, P. Yang, F. Fang, J. Xiang, C. C. Pain, I. M. Navon, Non-intrusive reduced order modelling of fluid– structure interactions, Computer Methods in Applied Mechanics and Engineering 303 (2016) 35–54

    work page 2016

  30. [30]

    X. Guo, W. Li, F. Iorio, Convolutional neural networks for steady flow approximation, in: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, 2016, pp. 481–490. 40

    work page 2016

  31. [31]

    Murata, K

    T. Murata, K. Fukami, K. Fukagata, Nonlinear mode decomposition with convolutional neural networks for fluid dynamics, Journal of Fluid Mechanics 882 (2020) A13

    work page 2020

  32. [32]

    Obiols-Sales, A

    O. Obiols-Sales, A. Vishnu, N. Malaya, A. Chandramowliswharan, Cfdnet: A deep learning-based accelerator for fluid simulations, in: Proceedings of the 34th ACM international conference on supercomputing, 2020, pp. 1–12

    work page 2020

  33. [33]

    Kochkov, J

    D. Kochkov, J. A. Smith, A. Alieva, Q. Wang, M. P. Brenner, S. Hoyer, Machine learning–accelerated computa- tional fluid dynamics, Proceedings of the National Academy of Sciences 118 (21) (2021) e2101784118

    work page 2021

  34. [34]

    Raissi, P

    M. Raissi, P. Perdikaris, G. E. Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computa- tional physics 378 (2019) 686–707

    work page 2019

  35. [35]

    Fan, J.-X

    X. Fan, J.-X. Wang, Differentiable hybrid neural modeling for fluid-structure interaction, Journal of Computa- tional Physics 496 (2024) 112584

    work page 2024

  36. [36]

    Benner, S

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

    work page 2015

  37. [37]

    M. A. Bouhlel, S. He, J. R. Martins, Scalable gradient–enhanced artificial neural networks for airfoil shape design in the subsonic and transonic regimes, Structural and Multidisciplinary Optimization 61 (4) (2020) 1363– 1376

    work page 2020

  38. [38]

    Y . S. Ong, P. B. Nair, A. J. Keane, Evolutionary optimization of computationally expensive problems via surro- gate modeling, AIAA journal 41 (4) (2003) 687–696

    work page 2003

  39. [39]

    Jeong, M

    S. Jeong, M. Murayama, K. Yamamoto, Efficient optimization design method using kriging model, Journal of aircraft 42 (2) (2005) 413–420

    work page 2005

  40. [40]

    R. B. Gramacy, H. K. Lee, Adaptive design and analysis of supercomputer experiments, Technometrics 51 (2) (2009) 130–145

    work page 2009

  41. [41]

    Sanchez, T

    R. Sanchez, T. Albring, R. Palacios, N. R. Gauger, T. D. Economon, J. J. Alonso, Coupled adjoint-based sen- sitivities in large-displacement fluid-structure interaction using algorithmic differentiation, International Journal for Numerical Methods in Engineering 113 (7) (2018) 1081–1107

    work page 2018

  42. [42]

    J. Hron, S. Turek, A monolithic fem/multigrid solver for an ale formulation of fluid-structure interaction with applications in biomechanics, in: Fluid-Structure Interaction: Modelling, Simulation, Optimisation, Springer, 2006, pp. 146–170. 41

    work page 2006

  43. [43]

    Colomés, F

    O. Colomés, F. Verdugo, I. Akkerman, A monolithic finite element formulation for the hydroelastic analysis of very large floating structures, International Journal for Numerical Methods in Engineering 124 (3) (2023) 714–751

    work page 2023

  44. [44]

    C. A. Felippa, K.-C. Park, C. Farhat, Partitioned analysis of coupled mechanical systems, Computer methods in applied mechanics and engineering 190 (24-25) (2001) 3247–3270

    work page 2001

  45. [45]

    , author Rallu, A

    C. Farhat, A. Rallu, K. Wang, T. Belytschko, Robust and provably second-order explicit-explicit and implicit- explicit staggered time-integrators for highly non-linear compressible fluid-structure interaction problems, Inter- national Journal for Numerical Methods in Engineering 84 (1) (2010) 73–107. doi:10.1002/nme.2883

    work page doi:10.1002/nme.2883 2010

  46. [46]

    X. Zhao, W. Ma, S. Islam, A. Narkhede, K. Wang, M2c: An open-source software for multiphysics simulation of compressible multi-material flows and fluid-structure interactions, Computer Physics Communications 321 (2026) 110023. doi:10.1016/j.cpc.2026.110023. URLhttps://doi.org

    work page doi:10.1016/j.cpc.2026.110023 2026

  47. [47]

    AeroS (2024). [link]. URLhttps://github.com/anarkh97/aero-s-foam

    work page 2024

  48. [48]

    M. P. Païdoussis, Fluid-Structure Interactions: Slender Structures and Axial Flow, V olume 1, 2nd Edition, Aca- demic Press/Elsevier, Amsterdam; Boston; Heidelberg; London; New York; Oxford, 2013, second edition, revised and updated

    work page 2013

  49. [49]

    C. K. Williams, C. E. Rasmussen, Gaussian processes for machine learning, V ol. 2, MIT press Cambridge, MA, 2006

    work page 2006

  50. [50]

    B. M. Adams, W. J. Bohnhoff, K. R. Dalbey, M. S. Ebeida, J. P. Eddy, M. S. Eldred, R. W. Hooper, P. D. Hough, K. T. Hu, J. D. Jakeman, et al., Dakota, a multilevel parallel object-oriented framework for design optimization, parameter estimation, uncertainty quantification, and sensitivity analysis: Version 6.16 user’s manual, Tech. rep., Sandia National L...

    work page doi:10.2172/1868142 2022

  51. [51]

    A. Main, X. Zeng, P. Avery, C. Farhat, An enhanced FIVER method for multi-material flow prob- lems with second-order convergence rate, Journal of Computational Physics 329 (2017) 141–172. doi:10.1016/j.jcp.2016.10.028

    work page doi:10.1016/j.jcp.2016.10.028 2017

  52. [52]

    Geuzaine, J.-F

    C. Geuzaine, J.-F. Remacle, Gmsh: A 3-d finite element mesh generator with built-in pre-and post-processing facilities, International journal for numerical methods in engineering 79 (11) (2009) 1309–1331

    work page 2009

  53. [53]

    Giordano, G

    J. Giordano, G. Jourdan, Y . Burtschell, M. Medale, D. Zeitoun, L. Houas, Shock wave impacts on deforming panel, an application of fluid-structure interaction, Shock Waves 14 (1) (2005) 103–110. 42

    work page 2005

  54. [54]

    Narkhede, N

    A. Narkhede, N. Eilers, E. Rivas, K. Wang, Multi-fidelity structural optimization for shock-driven fluid–structure interaction problems, in: AIAA SciTech Forum and Exposition, American Institute of Aeronautics and Astro- nautics, 2026. 43

    work page 2026