pith. sign in

arxiv: 2511.21031 · v2 · submitted 2025-11-26 · ⚛️ physics.flu-dyn

Hardware-Accelerated Phase-Averaging for Cavitating Bubbly Flows

Diego Vaca-Revelo , Benjamin Wilfong , Spencer H. Bryngelson , Aswin Gnanaskandan This is my paper

Pith reviewed 2026-05-17 05:24 UTC · model grok-4.3

classification ⚛️ physics.flu-dyn
keywords bubbly flowsphase averagingGPU accelerationcavitationmultiscale modelingKeller-Miksisacoustically driven flowsdilute suspensions
0
0 comments X

The pith

A GPU-accelerated phase-averaged solver accurately and efficiently simulates acoustically driven dilute bubbly flows.

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

This paper develops a multiscale approach to simulate bubbly suspensions driven by acoustic waves. The fluid is treated with compressible Navier-Stokes equations while bubbles are modeled at subgrid scale using either a Lagrangian volume-averaged or an ensemble-averaged formulation, both based on the Keller-Miksis equation. Validation against analytical solutions and experiments shows root-mean-square errors below 8 percent for bubble oscillations and collapse. On GPU hardware the method achieves a 16-fold speedup over a 64-core CPU, with good weak and strong scaling reported for both CPU and GPU runs.

Core claim

The proposed hardware-accelerated phase-averaged multiscale solver, using OpenACC for GPU offloading and subgrid bubble models via the Keller-Miksis equation, is robust, accurate, and efficient for simulating acoustically driven dilute bubbly flows, as evidenced by low-error validations and performance benchmarks showing substantial speedups and scalability.

What carries the argument

The phase-averaged multiscale solver that couples compressible Navier-Stokes for the carrier fluid with subgrid Lagrangian or ensemble-averaged bubble dynamics governed by the Keller-Miksis equation, accelerated via OpenACC directives on GPUs.

If this is right

  • The volume-averaged model allows detailed examination of individual bubble behaviors while matching experimental data.
  • The ensemble-averaged model further reduces computational expense by solving averaged equations instead of multiple realizations.
  • Good scalability is maintained across CPU and GPU platforms for both weak and strong scaling tests.
  • The approach supports multiscale simulations without resolving individual bubble interfaces.

Where Pith is reading between the lines

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

  • Similar acceleration strategies could apply to other dispersed-phase flow problems where subgrid models are appropriate.
  • Ensemble averaging may enable simulations at scales where individual bubble tracking becomes prohibitive.
  • Extensions to include weak bubble interactions could broaden applicability while retaining efficiency gains.

Load-bearing premise

The dilute suspension approximation and subgrid bubble models remain valid across the simulated acoustic driving conditions without requiring full resolution of bubble interfaces or strong bubble-bubble interactions.

What would settle it

A simulation or experiment in which strong bubble-bubble interactions occur or bubble interfaces must be fully resolved, showing significant deviation from the phase-averaged predictions.

Figures

Figures reproduced from arXiv: 2511.21031 by Aswin Gnanaskandan, Benjamin Wilfong, Diego Vaca-Revelo, Spencer H. Bryngelson.

Figure 1
Figure 1. Figure 1: Schematics of the (a) volume-averaged and (b) ensemble-averaged subgrid bubble models. the bubble contains both non-condensable gas and vapor, and we adopt the reduced-order models with constant heat and mass transfer coefficients at the bubble wall, as described by Preston et al. [36]. These models account for the effects of vapor and heat diffusion through the interface. Here, the vapor mass transfer rat… view at source ↗
Figure 2
Figure 2. Figure 2: General schematic of the Strang splitting algorithm for time integration. steppers remain below a tolerance of 10−4 . The algorithm alternates between advancing the states of the subgrid bubbles and the conservative variables of the flow field, ensuring consistent coupling between them. By manipulating the timescale requirements of the two systems, this strategy enables the efficient simulation of complex … view at source ↗
Figure 3
Figure 3. Figure 3: Evolution of an isolated bubble in response to a single cycle of a sinusoidal pressure wave using different CFL numbers and grid sizes. In the second validation case, we replicate the experimental observations reported by Ohl et al. [47], who investigated the dynamics of a trapped bubble in a water–glycerine mixture undergoing spherical collapse. The liquid host has a density of 1000 kg/m3 , viscosity of 0… view at source ↗
Figure 4
Figure 4. Figure 4: Radius evolution of the spherical collapse of an isolated bubble in response to a sinusoidal acoustic wave. Two test cases are considered for the bubble size distribution: a monodisperse case and a polydisperse case. In the monodisperse scenario, all bubbles have the same initial radius of 10 µm. In the polydisperse scenario, a more realistic distribution of bubble sizes is introduced by assigning radii ac… view at source ↗
Figure 5
Figure 5. Figure 5: Schematic of the dilute bubble screen configuration (not to scale). 13 [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Pressure profiles at the origin of the bubble screen using the volume-averaged and ensemble￾averaged subgrid models. (a) Monodisperse and (b) polydisperse bubble cloud. distance for the given void fraction and bubble radius is approximately 265 µm, which is more than twice the grid spacing. This satisfies the requirement of eq. (12). In the ensemble-averaged modeling framework, the polydispersity of bubble… view at source ↗
Figure 7
Figure 7. Figure 7: Computational cost of the volume-averaged (EL) and ensemble-averaged (EE) models on (a) CPU cores and (b) GPUs. cost of the EL model primarily depends on the total number of oscillating discrete bubbles in the domain and is independent of the bubble size distribution. Consequently, we restrict our EL tests to the monodisperse bubble cloud configuration. In contrast, the cost of the ensemble-averaged model … view at source ↗
Figure 8
Figure 8. Figure 8: Strong scaling on (a) AMD CPUs and (b) NVIDIA A100 GPUs for different problem sizes: (i) 16M, (ii) 32M, (iii) 64M grid cells. (a) AMD Milan CPU cores (b) NVIDIA A100 GPUs Grid Cells 16M 32M 64M 16M 32M 64M EL 74.21 % 78.64 % 85.31 % 29.69 % 42.70 % 57.68 % EE (monodisperse) 82.76 % 84.12 % 89.98 % 33.26 % 39.35 % 47.73 % EE (polydisperse) 92.43 % 89.19 % 94.02 % 38.13 % 46.52 % 52.77 % [PITH_FULL_IMAGE:fi… view at source ↗
Figure 9
Figure 9. Figure 9: Weak scaling on (a) AMD CPUs and (b) NVIDIA A100 GPUs for Euler–Lagrange (EL) and Euler–Euler (EE) models as labeled. 5.3.2. Weak scaling In weak scaling, the problem size assigned to each processor remains constant while the total number of compute devices increases. Within the EE framework, the number of bins in the polydisperse tests is held constant, and the problem size is therefore only determined by… view at source ↗
read the original abstract

We present a comprehensive validation, performance characterization, and scalability analysis of a hardware-accelerated phase-averaged multiscale solver designed to simulate acoustically driven dilute bubbly suspensions. The carrier fluid is modeled using the compressible Navier-Stokes equations. The dispersed phase is represented through two distinct subgrid formulations: a volume-averaged model that explicitly treats discrete bubbles within a Lagrangian framework, and an ensemble-averaged model that statistically represents the bubble population through a discretized distribution of bubble sizes. For both models, the bubble dynamics are modeled via the Keller--Miksis equation. For the GPU cases, we use OpenACC directives to offload computation to the GPUs. The volume-averaged model is validated against the analytical Keller-Miksis solution and experimental measurements, showing excellent agreement with root-mean-squared errors of less than 8% for both single-bubble oscillation and collapse scenarios. The ensemble-averaged model is validated by comparing it to volume-averaged simulations. On an NCSA Delta node with 4 NVIDIA A100 GPUs, we observe a speedup 16-fold compared to a 64-core AMD Milan CPU. The ensemble-averaged model offers additional reductions in computational cost by solving a single set of averaged equations, rather than multiple stochastic realizations. However, the volume-averaged model enables the interrogation of individual bubble dynamics, rather than the averaged statistics of the bubble dynamics. Weak and strong scaling tests demonstrate good scalability across both CPU and GPU platforms. These results show the proposed method is robust, accurate, and efficient for the multiscale simulation of acoustically driven dilute bubbly flows.

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 manuscript presents a hardware-accelerated phase-averaged multiscale solver for acoustically driven dilute bubbly suspensions. The carrier fluid is modeled with compressible Navier-Stokes equations while the dispersed phase uses Keller-Miksis bubble dynamics in either a volume-averaged Lagrangian formulation or an ensemble-averaged statistical formulation. Validation is reported for the volume-averaged model against analytical Keller-Miksis solutions and experiments (RMSE below 8% for single-bubble oscillation and collapse), with the ensemble-averaged model compared internally to volume-averaged runs. Performance results include a 16-fold GPU speedup on 4 NVIDIA A100s versus a 64-core CPU and good weak/strong scaling on both CPU and GPU platforms.

Significance. If the accuracy of the coupled compressible NS plus subgrid bubble solver holds for the full multiscale problem, the work would offer a practical route to efficient simulation of dilute cavitating flows by combining phase averaging with GPU offloading via OpenACC. The use of direct hardware timing and independent analytical/experimental references (rather than internally fitted quantities) strengthens the performance and speedup claims. However, the current validation scope limits the immediate significance for the integrated system under acoustic driving.

major comments (2)
  1. [Abstract and validation results] The central claim that the method is accurate for multiscale simulation of acoustically driven dilute bubbly flows rests on single-bubble validations (RMSE <8% vs analytical Keller-Miksis and experiments) plus internal consistency between volume- and ensemble-averaged models. No external reference solution or experimental benchmark is provided for the integrated system, including bubble-induced source terms in the fluid equations, local pressure feedback, or effects of phase averaging on nonlinear collapse dynamics under acoustic driving.
  2. [Validation of ensemble-averaged model] The ensemble-averaged model is validated solely by comparison to volume-averaged simulations; because the volume-averaged runs themselves lack an independent coupled-system benchmark, this does not establish accuracy of the phase-averaged equations for the target dilute suspensions.
minor comments (2)
  1. [Abstract] The abstract lacks details on the specific numerical schemes, time-stepping methods, error-bar reporting, and post-processing choices used to compute the reported RMSE values and speedups.
  2. [Model description] Clarify how the dilute-suspension assumption and subgrid bubble models are justified across the simulated acoustic driving amplitudes, particularly when bubble-induced pressure perturbations may become locally significant.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive review and for identifying areas where the validation discussion can be strengthened. We have revised the manuscript to clarify the scope of our benchmarks, add explicit discussion of limitations for the coupled system, and moderate claims in the abstract and conclusions. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Abstract and validation results] The central claim that the method is accurate for multiscale simulation of acoustically driven dilute bubbly flows rests on single-bubble validations (RMSE <8% vs analytical Keller-Miksis and experiments) plus internal consistency between volume- and ensemble-averaged models. No external reference solution or experimental benchmark is provided for the integrated system, including bubble-induced source terms in the fluid equations, local pressure feedback, or effects of phase averaging on nonlinear collapse dynamics under acoustic driving.

    Authors: We agree that independent experimental or numerical benchmarks for the fully coupled multiscale system under acoustic driving would further strengthen the work. Our validation strategy centers on rigorous, independent testing of the subgrid Keller-Miksis dynamics (against both analytical solutions and experiments) because these govern the bubble response that drives the source terms. The coupling itself follows standard volume-averaging procedures already established in the bubbly-flow literature. In the revised manuscript we have (i) added a new paragraph in Section 4 explicitly discussing the assumptions underlying the phase-averaged source terms and the dilute-limit regime in which they are expected to hold, (ii) inserted a forward-looking statement on the need for integrated benchmarks as future work, and (iii) revised the abstract and significance statements to emphasize that accuracy is demonstrated for the bubble dynamics and model consistency rather than claiming full-system experimental validation. These changes make the validation scope transparent while preserving the practical utility of the approach for dilute suspensions. revision: yes

  2. Referee: [Validation of ensemble-averaged model] The ensemble-averaged model is validated solely by comparison to volume-averaged simulations; because the volume-averaged runs themselves lack an independent coupled-system benchmark, this does not establish accuracy of the phase-averaged equations for the target dilute suspensions.

    Authors: We acknowledge that the ensemble-averaged validation is internal to our two formulations. However, the volume-averaged model itself has been validated against independent analytical Keller-Miksis solutions and experimental data for single-bubble oscillation and collapse (RMSE < 8 %). The ensemble-averaged equations are obtained by direct statistical averaging of the volume-averaged Lagrangian description in the dilute limit; therefore the comparison quantifies the fidelity of the discretization and averaging procedure rather than introducing a new source of error. In the revised manuscript we have expanded the derivation in Section 2.3 to show the explicit relationship between the two models and added a short discussion of the expected error bounds for dilute bubbly suspensions. We have also noted that full external validation of the integrated system is an important open direction. These additions address the referee’s concern without overstating the current evidence. revision: partial

Circularity Check

0 steps flagged

No circularity: accuracy and performance claims rest on independent external benchmarks and direct hardware measurements

full rationale

The paper validates the volume-averaged model directly against the analytical Keller-Miksis solution and experimental measurements (RMSE <8% for oscillation and collapse), while the ensemble-averaged model is checked for internal consistency against volume-averaged runs. Speedup factors (16x on 4 A100 GPUs vs 64-core CPU) are obtained from explicit timing measurements. The underlying equations (compressible Navier-Stokes plus Keller-Miksis) are standard and not redefined in terms of the target outputs. No fitted parameters are repurposed as predictions, no self-citation chain supplies a uniqueness theorem or ansatz, and no renaming of known results occurs. All load-bearing claims are therefore supported by external references rather than reducing to the paper's own inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claims rest on standard compressible Navier-Stokes and Keller-Miksis models drawn from prior literature plus the assumption that OpenACC offloading preserves numerical fidelity; no new free parameters or invented entities are introduced in the abstract.

axioms (2)
  • domain assumption Keller-Miksis equation accurately captures bubble radial dynamics under the dilute acoustic driving conditions considered.
    Invoked for both volume- and ensemble-averaged formulations without additional derivation or sensitivity analysis shown in the abstract.
  • domain assumption OpenACC directives correctly map the computation to GPUs without introducing floating-point or ordering artifacts that alter the reported speedups or errors.
    Required for the performance characterization on A100 GPUs versus CPU baseline.

pith-pipeline@v0.9.0 · 5592 in / 1383 out tokens · 93869 ms · 2026-05-17T05:24:16.689662+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

48 extracted references · 48 canonical work pages

  1. [1]

    Neppiras, Acoustic cavitation, Physics Reports61(1980) 159–251

    E. Neppiras, Acoustic cavitation, Physics Reports61(1980) 159–251

    work page 1980

  2. [2]

    Yasui, Acoustic cavitation, in: Acoustic Cavitation and Bubble Dynamics, Springer, 2017, pp

    K. Yasui, Acoustic cavitation, in: Acoustic Cavitation and Bubble Dynamics, Springer, 2017, pp. 1–35

    work page 2017

  3. [3]

    E. B. Flint, K. S. Suslick, The temperature of cavitation, Science253(1991) 1397–1399

    work page 1991

  4. [4]

    Lauterborn, T

    W. Lauterborn, T. Kurz, Physics of bubble oscillations, Reports on Progress in Physics73 (2010) 106501

    work page 2010

  5. [5]

    S. Beig, B. Aboulhasanzadeh, E. Johnsen, Temperatures produced by inertially collapsing bubbles near rigid surfaces, Journal of Fluid Mechanics852(2018) 105–125

    work page 2018

  6. [6]

    Ikeda, S

    T. Ikeda, S. Yoshizawa, N. Koizumi, M. Mitsuishi, Y. Matsumoto, Focused ultrasound and lithotripsy, Therapeutic Ultrasound (2016) 113–129

    work page 2016

  7. [7]

    Kooiman, S

    K. Kooiman, S. Roovers, S. A. Langeveld, R. T. Kleven, H. Dewitte, M. A. O’Reilly, J.-M. Escoffre, A. Bouakaz, M. D. Verweij, K. Hynynen, et al., Ultrasound-responsive cavitation nuclei for therapy and drug delivery, Ultrasound in Medicine & Biology46(2020) 1296–1325

    work page 2020

  8. [8]

    C. C. Coussios, R. A. Roy, Applications of acoustics and cavitation to noninvasive therapy and drug delivery, Annual Review of Fluid Mechanics40(2008) 395–420

    work page 2008

  9. [9]

    J. H. Bang, K. S. Suslick, Applications of ultrasound to the synthesis of nanostructured materials, Advanced Materials22(2010) 1039–1059

    work page 2010

  10. [10]

    J. Luo, Z. Fang, R. L. Smith Jr, Ultrasound-enhanced conversion of biomass to biofuels, Progress in Energy and Combustion Science41(2014) 56–93

    work page 2014

  11. [11]

    K. S. Suslick, Y. Didenko, M. M. Fang, T. Hyeon, K. J. Kolbeck, W. B. McNamara III, M. M. Mdleleni, M. Wong, Acoustic cavitation and its chemical consequences, Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences357(1999) 335–353

    work page 1999

  12. [12]

    Voronin, G

    D. Voronin, G. Sankin, V. Teslenko, R. Mettin, W. Lauterborn, Secondary acoustic waves in a polydisperse bubbly medium, Journal of Applied Mechanics and Technical Physics44(2003) 17–26

    work page 2003

  13. [13]

    Leroy, A

    V. Leroy, A. Strybulevych, J. H. Page, M. G. Scanlon, Sound velocity and attenuation in bubbly gels measured by transmission experiments, The Journal of the Acoustical Society of America123(2008) 1931–1940

    work page 2008

  14. [14]

    Kaneko, T

    Y. Kaneko, T. Maruyama, K. Takegami, T. Watanabe, H. Mitsui, K. Hanajiri, H. Nagawa, Y. Matsumoto, Use of a microbubble agent to increase the effects of high intensity focused ultrasound on liver tissue, European Radiology15(2005) 1415–1420

    work page 2005

  15. [15]

    Kajiyama, K

    K. Kajiyama, K. Yoshinaka, S. Takagi, Y. Matsumoto, Micro-bubble enhanced HIFU, Physics Procedia3(2010) 305–314

    work page 2010

  16. [16]

    D. J. Chung, S. H. Cho, J. M. Lee, S.-T. Hahn, Effect of microbubble contrast agent during high intensity focused ultrasound ablation on rabbit liver in vivo, European Journal of Radiology 81(2012) e519–e523. 21

    work page 2012

  17. [17]

    E. K. Juang, L. H. De Koninck, K. S. Vuong, A. Gnanaskandan, C.-T. Hsiao, M. A. Averkiou, Controlled hyperthermia with high-intensity focused ultrasound and ultrasound contrast agent microbubbles in porcine liver, Ultrasound in Medicine & Biology49(2023) 1852–1860

    work page 2023

  18. [18]

    Canselier, H

    J. Canselier, H. Delmas, A. Wilhelm, B. Abismail, Ultrasound emulsification—An overview, Journal of Dispersion Science and Technology23(2002) 333–349

    work page 2002

  19. [19]

    Krasulya, V

    O. Krasulya, V. Bogush, V. Trishina, I. Potoroko, S. Khmelev, P. Sivashanmugam, S. Anandan, Impact of acoustic cavitation on food emulsions, Ultrasonics Sonochemistry30(2016) 98–102

    work page 2016

  20. [20]

    Dular, T

    M. Dular, T. Griessler-Bulc, I. Gutierrez-Aguirre, E. Heath, T. Kosjek, A. K. Klemenˇ ciˇ c, M. Oder, M. Petkovˇ sek, N. Raˇ cki, M. Ravnikar, et al., Use of hydrodynamic cavitation in (waste) water treatment, Ultrasonics Sonochemistry29(2016) 577–588

    work page 2016

  21. [21]

    Zhang, A

    D. Zhang, A. Prosperetti, Ensemble phase-averaged equations for bubbly flows, Physics of Fluids6(1994) 2956–2970

    work page 1994

  22. [22]

    K. Ando, T. Colonius, C. E. Brennen, Numerical simulation of shock propagation in a polydisperse bubbly liquid, International Journal of Multiphase Flow37(2011) 596–608

    work page 2011

  23. [23]

    Maeda, T

    K. Maeda, T. Colonius, Eulerian–Lagrangian method for simulation of cloud cavitation, Journal of Computational Physics371(2018) 994–1017

    work page 2018

  24. [24]

    Gnanaskandan, C.-T

    A. Gnanaskandan, C.-T. Hsiao, G. Chahine, Modeling of microbubble-enhanced high-intensity focused ultrasound, Ultrasound in Medicine & Biology45(2019) 1743–1761

    work page 2019

  25. [25]

    S. H. Bryngelson, K. Schmidmayer, T. Colonius, A quantitative comparison of phase-averaged models for bubbly, cavitating flows, International Journal of Multiphase Flow115(2019) 137–143

    work page 2019

  26. [26]

    Snir, MPI–The Complete Reference: The MPI core, volume 1, MIT Press, 1998

    M. Snir, MPI–The Complete Reference: The MPI core, volume 1, MIT Press, 1998

    work page 1998

  27. [27]

    B¨ ohme, Characterizing load and communication imbalance in parallel applications, volume 23, Forschungszentrum J¨ ulich, 2014

    D. B¨ ohme, Characterizing load and communication imbalance in parallel applications, volume 23, Forschungszentrum J¨ ulich, 2014

    work page 2014

  28. [28]

    P. Wang, T. Abel, R. Kaehler, Adaptive mesh fluid simulations on gpu, New Astronomy15 (2010) 581–589

    work page 2010

  29. [29]

    Salvadore, M

    F. Salvadore, M. Bernardini, M. Botti, Gpu accelerated flow solver for direct numerical simulation of turbulent flows, Journal of Computational Physics235(2013) 129–142

    work page 2013

  30. [30]

    Sweet, D

    J. Sweet, D. H. Richter, D. Thain, GPU acceleration of Eulerian–Lagrangian particle-laden turbulent flow simulations, International Journal of Multiphase Flow99(2018) 437–445

    work page 2018

  31. [31]

    Radhakrishnan, H

    A. Radhakrishnan, H. Le Berre, B. Wilfong, J.-S. Spratt, M. Rodriguez Jr., T. Colonius, S. H. Bryngelson, Method for portable, scalable, and performant GPU-accelerated simulation of multiphase compressible flow, Computer Physics Communications302(2024) 109238

    work page 2024

  32. [32]

    Piscaglia, F

    F. Piscaglia, F. Ghioldi, GPU acceleration of CFD simulations in OpenFOAM, Aerospace10 (2023) 792

    work page 2023

  33. [33]

    D. C. Jespersen, Acceleration of a CFD code with a GPU, Scientific Programming18(2010) 193–201. 22

    work page 2010

  34. [34]

    S. H. Bryngelson, K. Schmidmayer, V. Coralic, J. C. Meng, K. Maeda, T. Colonius, MFC: An open-source high-order multi-component, multi-phase, and multi-scale compressible flow solver, Computer Physics Communications266(2021) 107396

    work page 2021

  35. [35]

    Wilfong, H

    B. Wilfong, H. Le Berre, A. Radhakrishnan, A. Gupta, D. Vaca-Revelo, D. Adam, H. Yu, H. Lee, J. R. Chreim, M. Carcana Barbosa, Y. Zhang, E. Cisneros-Garibay, A. Gnanaskandan, M. Rodriguez Jr., R. D. Budiardja, S. Abbott, T. Colonius, S. H. Bryngelson, MFC 5.0: An exascale many-physics flow solver, arXiv preprint arXiv:2503.07953 (2025)

    work page arXiv 2025

  36. [36]

    Preston, T

    A. Preston, T. Colonius, C. Brennen, A reduced-order model of diffusive effects on the dynamics of bubbles, Physics of Fluids19(2007)

    work page 2007

  37. [37]

    Maeda, T

    K. Maeda, T. Colonius, A source term approach for generation of one-way acoustic waves in the Euler and Navier–Stokes equations, Wave Motion75(2017) 36–49

    work page 2017

  38. [38]

    S. H. Bryngelson, Fast integration method for averaging polydisperse bubble population dynamics, Computers & Fluids304(2026) 106877

    work page 2026

  39. [39]

    Sinha, S

    A. Sinha, S. H. Bryngelson, Neural networks can be FLOP-efficient integrators of 1D oscillatory integrands, Transactions on Machine Learning Research (2024)

    work page 2024

  40. [40]

    S. H. Bryngelson, R. O. Fox, T. Colonius, Conditional moment methods for polydisperse cavitating flows, Journal of Computational Physics477(2023) 111917

    work page 2023

  41. [41]

    Charalampopoulos, S

    A. Charalampopoulos, S. H. Bryngelson, T. Colonius, T. P. Sapsis, Hybrid quadrature moment method for accurate and stable representation of non-Gaussian processes and their dynamics, Philosophical Transactions of the Royal Society A380(2022)

    work page 2022

  42. [42]

    Colonius, R

    T. Colonius, R. Hagmeijer, K. Ando, C. E. Brennen, Statistical equilibrium of bubble oscillations in dilute bubbly flows, Physics of Fluids20(2008)

    work page 2008

  43. [43]

    Gottlieb, C.-W

    S. Gottlieb, C.-W. Shu, Total variation diminishing Runge-Kutta schemes, Mathematics of Computation67(1998) 73–85

    work page 1998

  44. [44]

    Strang, On the construction and comparison of difference schemes, SIAM Journal on Numerical Analysis5(1968) 506–517

    G. Strang, On the construction and comparison of difference schemes, SIAM Journal on Numerical Analysis5(1968) 506–517

    work page 1968

  45. [45]

    Chandrasekaran, G

    S. Chandrasekaran, G. Juckeland, OpenACC for programmers: Concepts and strategies, Addison-Wesley Professional, 2017

    work page 2017

  46. [46]

    K. W. Thompson, Time dependent boundary conditions for hyperbolic systems, Journal of Computational Physics68(1987) 1–24

    work page 1987

  47. [47]

    C.-D. Ohl, T. Kurz, R. Geisler, O. Lindau, W. Lauterborn, Bubble dynamics, shock waves and sonoluminescence, Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences357(1999) 269–294

    work page 1999

  48. [48]

    URL:https://docs.ncsa.illinois.edu

    NCSA, Delta Architecture User Guide, 2023. URL:https://docs.ncsa.illinois.edu. 23

    work page 2023