Pith. sign in

REVIEW 3 major objections 5 minor 36 references

Temperature and pressure reconstruction in turbulent Rayleigh-B\'enard convection by Lagrangian velocities using PINN

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A physics-informed neural network reconstructs temperature and pressure from Lagrangian velocity data in turbulent Rayleigh-Bénard convection, matching DNS ground truth at 90% correlation.

desk verdict A credible extension of PINN-based temperature/pressure reconstruction to hard-turbulence RBC, but the DNS-calibrated mean profile makes the headline validation less independent than it looks. read the letter →

arxiv 2505.02580 v2 pith:CMVBVCWT submitted 2025-05-05 physics.flu-dyn physics.comp-ph

R. Barta , M.-C. Volk , C. Bauer , C. Wagner , M. Mommert This is my paper · ORCID
classification physics.flu-dynphysics.comp-ph
keywords physics-informedneuralnetworksRayleigh-BénardconvectionparticletrackingvelocimetrytemperaturereconstructionpressureLagrangianvelocitydataassimilationturbulent
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that a single physics-informed neural network, fed only Lagrangian velocity measurements from tracer particles, can reconstruct the temperature and pressure fields of turbulent Rayleigh-Bénard convection at $Ra = 10^9$ and $Pr \approx 7$. If true, it would give experimentalists a way to obtain all three main flow variables from particle tracking alone, without separate thermometry or pressure probes. The claim is demonstrated two ways: against DNS ground truth, where reconstructed fields reach about 90% correlation, and on experimental PTV data, where the network produces physically realistic plumes, large-scale circulation, and pressure extrema while also smoothing noisy velocity components.

What carries the argument

The central object is a modified PINN output: the MLP emits a temperature fluctuation $T'$ and pressure $p$, and the true temperature is assembled as $T = T_{\mathrm{mean}}(Z,a) + (Z^2 - Z)T'$, where $T_{\mathrm{mean}}(Z,a)$ is an empirical two-sided exponential vertical profile. This form enforces the top and bottom temperature boundary conditions exactly and converts the hard part of the problem, steep wall gradients and a well-mixed bulk, into a fluctuation-learning task. The parameter $a$ is fixed by the Nusselt number through $Nu = a/(2 - 2e^{-0.5a}) \approx a/2$, set to $a = 126$ from the DNS value $Nu \approx 63$. The rest of the machinery is a standard equation-constrained MLP training loop: Adam optimization of a loss combining velocity data, momentum and energy and divergence residuals, pressure centering, and velocity boundary conditions, with sine activation functions that help the network represent many wavenumbers.

What would settle it

Place a vertical rake of fine thermistors in the experimental cell, time-average their readings over many large-scale-circulation periods, and compare the resulting mean temperature profile with $T_{\mathrm{mean}}(Z, a=126)$. If the measured profile deviates beyond boundary-layer uncertainty, the reconstructed temperature field and derived heat-transport number are set by the assumed prior rather than by the data.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that hard-turbulence Rayleigh-Bénard convection ($Ra = 10^9$, $Pr \approx 7$) is no longer out of reach for PINN-based assimilation: once the network is tasked with learning only temperature fluctuations around an empirical mean profile instead of the full temperature field, a single multilayer perceptron with sine activations can infer both temperature and pressure from Lagrangian velocity data. In the DNS validation this yields Pearson correlations around 0.90 for temperature and pressure at particle positions, with mean absolute errors near $10^{-2}$ for temperature and $10^{-3}$ for pressure. On experimental PTV data with the same nominal parameters, the same recipe produces thermal plumes, a large-scale circulation, and pressure extrema at the expected locations, and it visibly smooths measurement noise in the camera-direction velocity component.

Load-bearing premise

The reconstruction leans on an assumed mean temperature profile calibrated to the simulation's heat-transport level remaining valid in the experimental cell, whose glass sidewalls lose about 20 percent of the heat; the much lower heat-transport estimate the network produces for the experiment (34.6 versus the assumed 63) suggests the premise is already strained.

Editorial extensions

If this is right

  • With only tracer-particle velocities spanning about 2.5 free-fall times, the network reproduces temperature and pressure structures at roughly 90% correlation with DNS ground truth at particle positions.
  • Experimental PTV data can be upgraded from velocity-only to velocity-temperature-pressure fields, revealing thermal plumes, the large-scale circulation, and pressure minima in corner and central circulations.
  • The equation-constrained velocity output acts as a denoiser, particularly along the camera axis where PTV errors are largest.
  • The reconstructed flow can be used to estimate the Nusselt number, yielding 61.8 for DNS (close to 63.4) and a heat-loss-reduced value for the experiment that rises to about 55 after the authors' 1.6 velocity-scale correction.
  • Since the network is released as open-source code and plugs into the tracking workflow used for the experiment, the pipeline can be applied to other convection experiments without new hardware.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: because $a$ is fixed from DNS rather than learned, the method is not fully data-driven; treating $a$ as a trainable parameter or inferring it from sparse temperature measurements would make the pipeline self-calibrating in setups with unknown heat loss.
  • Editorial inference: the 90% correlation is scored on particle positions, where boundary layers are absent; the paper's own error maps show the largest deviations near the heating and cooling plates, so an Eulerian grid-based validation including boundary layers would likely report a lower correlation.
  • Editorial inference: the same mean-profile trick should transfer to other wall-bounded convective flows whenever a reliable analytic mean-temperature form exists; flows without such a form would need a different boundary-condition prior.
  • Editorial inference: combining Lagrangian velocity with sparse point thermometry, such as a few thermistors or thermochromic particles, would provide a direct experimental check and could extend the method's accuracy beyond the reported bulk-focused metrics.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper presents a physics-informed neural network (PINN) method to reconstruct temperature and pressure fields from Lagrangian velocity data in turbulent Rayleigh-Bénard convection at Ra=10^9 and Pr=6.9. The network, a multilayer perceptron with sine activation, is trained on DNS-generated particle tracks (150,000 particles) and on experimental PTV tracks (about 50,000 active particles per time step). A key ingredient is an empirical mean temperature profile Tmean(Z,a) with parameter a set to 2Nu, which imposes the wall boundary conditions and reduces the difficulty of the near-wall gradients. In the DNS validation, the reconstructed temperature and pressure reach Pearson correlations of about 90% and mean absolute errors around 10^-2 and 10^-3, respectively; the reconstructed Nusselt number is 61.8 compared with 63.4 in the DNS. For the experiment, the same a=126 is used, yielding a Nusselt number of 34.6, later corrected to 55.3 using a velocity-scale factor. The paper claims the method can provide temperature and pressure fields in the hard-turbulence regime from PTV data alone, and it makes the PINN code available open-source. I also checked the internal consistency of Eq. (7): the printed mean-temperature profile does satisfy Tmean(0)=+0.5, Tmean(1)=-0.5, and yields Eq. (8) by differentiating at Z=0, so the stress-test concern about Eq. (7) does not land.

Significance. If the result holds, the method would be a practical step toward measuring all three key fields (velocity, temperature, pressure) in turbulent thermal convection using only particle-tracked velocity data. The DNS validation is quantitative, with high correlation on temperature and pressure, and the open-source code is a strength for reproducibility. The use of a prescribed mean profile is a reasonable way to overcome the hard-turbulence difficulty, and the experimental demonstration is a useful proof of concept, though it lacks ground-truth temperature or pressure data. The main value of the paper lies in showing that the approach works in a regime (Ra=10^9) that previously had only been demonstrated at lower Rayleigh numbers.

major comments (3)
  1. [Section 4, evaluation metrics (Fig. 8)] The PCC and MAE values are computed on the last 10 time steps of the DNS dataset, which are part of the 125 time steps used for training the PINN through the data-loss term. Because the manuscript does not describe a held-out validation partition, the 90% correlation does not establish that the PINN can reconstruct temperature and pressure from velocity data outside the training interval. Please either train on the first 115 time steps and evaluate on the final 10, or explicitly state that the reported correlation is a training-fidelity metric rather than a generalization metric.
  2. [Section 4, Nusselt-number validation] The mean temperature profile parameter a is set to 2Nu with Nu=63.4 taken from the DNS, so the reported Nu_PINN=61.8 is largely constrained by the prescribed mean profile and is not an independent physical validation. The manuscript should present this as a consistency check. Additionally, reporting the correlation and MAE for the temperature fluctuation field T' alone, which is the part actually learned by the network, would strengthen the validation and clarify how much of the 90% correlation is attributable to the prescribed mean profile.
  3. [Section 5, experimental reconstruction] In the experimental case there is no measured temperature or pressure ground truth, and the mean temperature profile is prescribed with a=126 based on the DNS Nusselt number. The reconstructed Nu_PINN=34.6 deviates strongly from the assumed value of 63, and the correction factor of 1.6 is derived solely from the ratio of maximum vertical velocities, not from an independent heat-flux measurement. The manuscript should either obtain or cite an experimental Nusselt number for the cell, or explicitly label the experimental temperature and pressure fields as model-dependent reconstructions whose accuracy is not yet validated.
minor comments (5)
  1. [Section 2.2, paragraph on track visualization] The text states the maximum observed velocity is about ±10 m/s, which is inconsistent with the later statement that the maximum normalized vertical velocity is about 0.18 in the experiment (with a reference velocity of about 50 mm/s). The units appear to be a typo and should be corrected (e.g., ±0.10 m/s or ±10 cm/s).
  2. [Section 3.1, discussion of Eq. (7)] The statement that a=1 results in a linear temperature profile is only approximately true; the profile is exponential, and for a=1 the deviations from a straight line connecting the boundary values are small but visible. Consider rephrasing to 'approximately linear' or 'close to linear'.
  3. [Section 3.1, Eq. (6)] The factor (Z^2 - Z) multiplying the temperature fluctuation is introduced without explanation. It would be helpful to state explicitly that this factor forces the fluctuation to vanish at the top and bottom plates, so that the boundary conditions are exactly satisfied by Tmean alone.
  4. [Figure 7 caption] The caption says the profile is 'predicted in the DNS', but the figure shows the empirical function Tmean(Z,a) for increasing a. Clarify that the DNS data are used to motivate the value a=126, while the curves are from Eq. (7).
  5. [References] The reference to Barta et al. (2023) is marked 'under review'; if a published version is available, it should be updated.

Circularity Check

2 steps flagged · score 6.0 of 10

Nusselt validation reduces to the a=2Nu mean-profile calibration, and part of the 90% temperature correlation inherits the prescribed profile; pressure and fluctuation fields remain independent.

  1. fitted input called prediction [Sec. 3.1, Eq. (8); Sec. 4, Nusselt verification paragraph]
    "The physical meaning of a is derived by evaluating equation (5) on the heating or cooling plate where the velocity is zero, and using the derivative of equation (6) with respect to Z. Thus, we can relate the parameter a with the Nusselt number: Nu = a/(2−2e^{−0.5a}) ≃ a/2, if a≫1, and set a = 2 Nu = 126 throughout the paper, because we know that the average Nusselt number is Nu = 63 obtained from DNS in our case, see section 2.1. ..."

    The mean profile Tmean(Z,a) in Eqs. (6)-(7) fixes the mean vertical temperature gradient and, through the wall values, the dominant Nusselt-number integrand of Eq. (5). Setting a = 2Nu = 126 from the DNS makes Nu ≈ a/2 by Eq. (8) by construction. Recomputing Nu from the reconstructed temperature field and reporting Nu_PINN = 61.8 as validation is therefore a consistency check of the calibration, not an independent inference. The learned fluctuation field and pressure are independent content, but this Nusselt-number agreement is forced by the input value.

  2. self definitional [Sec. 3.1, Eq. (6); Sec. 4, correlation results]
    "Equation (6) is chosen empirically and ensures that the temperature boundary conditions are satisfied exactly. ... The parameter a = 126 is chosen to mimic the mean temperature profile and the temperature boundary conditions at the top and bottom plates for the considered case with a Nusselt number Nu = 63. ... the reconstructed temperature and pressure fields also reach correlation values of about 90% with respect to the ground truth at the end of training."

    Because the network output is defined as T = Tmean(Z,a) + (Z^2 − Z)T′, the vertical mean profile and wall temperatures are injected from the same DNS that supplies Nu = 63 used to set a. The reported 90% temperature correlation is therefore partly a measure of how well the prescribed mean profile fits the DNS mean, rather than purely a test of the learned reconstruction. The fluctuation field T′ is still inferred from velocity data, so the circularity is partial; the pressure correlation is independent of this construction.

full rationale

The reconstruction pipeline is not entirely circular: the PINN genuinely learns pressure and temperature fluctuations from the Lagrangian velocity data, and the pressure field is not tied to the a calibration. However, in the DNS validation the mean vertical temperature profile and its wall gradient are prescribed from the DNS Nusselt number through a = 2Nu = 126, and Eq. (8) makes Nu ≈ a/2 by construction. The paper's Nusselt verification, Nu_PINN = 61.8 versus DNS Nu = 63, is therefore a consistency check of that input calibration rather than an independent prediction. The 90% temperature correlation also inherits this prescribed mean profile, although the learned fluctuation component is genuine. The experimental section transfers the same a from DNS to the PTV cell and later attributes the resulting Nu_PINN = 34.6 to sidewall heat loss; that is an external transfer assumption rather than a circular step, but it inherits the same calibration. Self-citations to Mommert et al. (2024) and Volk et al. (2025) supply architecture, loss weights, and evaluation practices, but the equations in this paper state the mechanism and are not themselves the source of the partial circularity. Overall, the central temperature/Nusselt validation is partially forced by the fitted mean-temperature profile, while the pressure reconstruction and the fluctuation field retain independent content.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central result rests on the Oberbeck-Boussinesq model and on a hand-designed mean temperature profile whose parameter a is taken from a separate DNS Nusselt number. The experimental application further assumes the DNS-based profile is transferable despite non-adiabatic sidewalls. No new physical entities are introduced.

free parameters (3)
  • a (mean temperature profile parameter) = 126 (a=2Nu with Nu=63 from DNS)
    Sets the shape of the prescribed mean temperature profile in Eq. (7) and the wall temperature gradient. The value is taken from the DNS Nusselt number, not inferred from the particle data.
  • Sidewall heat-loss correction factor = 1.6
    Applied post hoc to the experimental Nusselt number (Nu=34.6 x 1.6 = 55.3) to account for non-adiabatic glass sidewalls. It is chosen as the ratio of DNS to experimental maximum vertical velocity and is not derived from the PINN itself.
  • Loss weights lambda_data, lambda_NS, lambda_EE, lambda_div, lambda_C, lambda_BC = 1.0, 0.1, 0.01, 0.001, 0.001, 0.0001
    Hand-tuned weights in Eq. (9) following Mommert et al. (2024); they affect the balance between data fitting and physical residuals but are not fitted to the validation metrics.
assumptions (5)
  • domain assumption The Oberbeck-Boussinesq equations (1)-(3) with incompressibility govern the flow.
    Standard model for Rayleigh-Bénard convection; validity for the experiment is asserted via Gray and Giorgini (1976) and Weiss et al. (2024).
  • ad hoc to paper The empirical mean temperature profile Tmean(Z,a) in Eq. (6)-(7) is an adequate representation of the true mean field.
    The functional form is chosen empirically and is not derived from the governing equations; its parameter a is fixed by the DNS Nusselt number.
  • ad hoc to paper The experimental Nusselt number is approximately 63, the same as the DNS, so a=126 is used for the experiment.
    Section 5 states the same a is chosen for the experiment because Nu is known from DNS for similar parameters, despite non-adiabatic sidewalls.
  • domain assumption No-slip and impermeable boundary conditions apply at all walls.
    Used in the boundary loss term LBC (Eq. 15) and in the DNS; standard for solid walls.
  • domain assumption The PINN loss landscape is smooth enough that minimizing the weighted residuals (Eq. 9) yields the physical solution rather than a spurious local minimum.
    Implicit in all PINN applications; no convergence guarantees are proven.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Temperature and pressure reconstruction in turbulent Rayleigh-B\'enard convection by Lagrangian velocities using PINN." pith.science (2026). https://pith.science/paper/CMVBVCWT

@misc{pith2026250502580,
  author       = {Pith},
  title        = {Pith review of: Temperature and pressure reconstruction in turbulent Rayleigh-B\'enard convection by Lagrangian velocities using PINN},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CMVBVCWT}},
  note         = {Machine review of arXiv:2505.02580}
}
abstract

Velocity, pressure, and temperature are the key variables for understanding thermal convection, and measuring them all is a complex task. In this paper, we demonstrate a method to reconstruct temperature and pressure fields based on given Lagrangian velocity data. A physics-informed neural network (PINN) based on a multilayer perceptron architecture and a periodic sine activation function is used to reconstruct both the temperature and the pressure for two cases of turbulent Rayleigh-B\'enard convection (Pr = 6.9, Ra = $10^9$). The first dataset is generated with DNS and it includes Lagrangian velocity data of 150000 tracer particles. The second contains a PTV experiment with the same system parameters in a water-filled cubic cell, and we observed about 50000 active particle tracks per time step with the open-source framework proPTV. A realistic temperature and pressure field could be reconstructed in both cases, which underlines the importance of PINNs also in the context of experimental data. In the case of the DNS, the reconstructed temperature and pressure fields show a 90\% correlation over all particles when directly validated against the ground truth. Thus, the proposed method, in combination with particle tracking velocimetry, is able to provide velocity, temperature, and pressure fields in convective flows even in the hard turbulence regime. The PINN used in this paper is compatible with proPTV and is part of an open source project. It is available on request at https://github.com/DLR-AS-BOA.

Figures

Figures reproduced from arXiv: 2505.02580 by the authors.

Figure 1
Figure 1. All tracks generated with DNS colored by their vertical velocity. 2.2. PTV experiment The PTV experiment is the same water-filled cubic Rayleigh-B´enard convection cell as discussed in Barta and Wagner (2025), except here we use a temperature gradient between the heating and cooling plate of ∆T ≈ 4 K, and a recoding frequency of f = 10 Hz. For the temperature difference considered, the Oberbeck-Boussinesq approximat… view at source ↗
Figure 2
Figure 2. Technical drawing of the Rayleigh-B´enard experiment. A cubic cell filled with water and heated from below and cooled from above, four cameras observing the flow and two LED arrays illuminating the flow. The green arrows indicate the reference coordinate system [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. provides an overview of the number of triangulated particles, initialized tracks, active tracks, and broken tracks during PTV processing of the 150 time steps [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (13 more)
Figure 4
Figure 4. Figure 4: All PTV tracks colored by their normalized vertical velocity. In figure 5 the track length histogram is shown and an average track length of 26 is calculated. In the experiment, the velocity field has full temporal resolution. This is verified because the smallest velo…
Figure 5
Figure 5. Figure 5: Track length histogram. The average track length is about 26. 3. PINN Methodology 3.1. Architecture A multi-layer perceptron (MLP) is used to reconstruct unknown flow fields, such as temperature and pressure fields from known velocity fields. The MLP, shown in figure 6…
Figure 6
Figure 6. Figure 6: Architecture of the MLP used as physics-informed neural network to assimilate temperature and pressure with NN neurons per hidden layer, NL hidden layers, and a sinus activation function: σ(·) = sin(·). The network is an extension of the method introduced by Mommert et…
Figure 7
Figure 7. Figure 7: The mean temperature profile as a function of the vertical cell direction Z averaged with respect to time and the horizontal plane defined by the coordinates X and Y predicted in the DNS for Ra = 109 and Pr = 6.9. The empirical function Tmean(Z, a) is shown as dashed l…
Figure 8
Figure 8. Figure 8: Pearson correlation values (left) and MAE (right) of the velocity components, temperature and pressure estimated from the reconstructed PINN fields compared with the DNS fields over 2500 epochs of training [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: Colored Instantaneous vertical velocity (top), temperature (mid) and pressure (bottom) at the time step 120. Left: DNS vertical velocity, temperature and pressure on the particle positions. Right: reconstructed vertical velocity, temperature and pressure on the particl…
Figure 10
Figure 10. Figure 10: Computed and reconstructed mean temperature profiles over the cell height determined from the temperatures of the particles in the DNS and the PINN, respectively [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: PDFs of the three velocity components, temperature, pressure and heat flux calculated from the synthetic PTV tracks from the DNS in green compared with the PDFs obtained with the PINN results in red. Compared to the overall picture of the flow given in figure 9, a dir…
Figure 12
Figure 12. Figure 12: Color contours of the three velocity components in the diagonal plane of the cubic cell containing the LSC with a thickness of L/50 predicted in the DNS uDNS, vDNS, wDNS (ground truth) in comparison with the reconstructed velocities u, v, w computed from the synthetic…
Figure 13
Figure 13. Figure 13: Color contours of the pressure and temperature in the diagonal plane of the cubic cell containing the LSC with a thickness of L/50 predicted in the DNS TDNS, pDNS (ground truth) in comparison with the reconstructed properties T, p computed from the synthetic PTV track…
Figure 14
Figure 14. Figure 14: Color contours of the three velocity components in the diagonal plane of the cubic cell containing the LSC with a thickness of L/20 measured with PTV experimentally uEXP , vEXP , wEXP in comparison with the reconstructed velocities u, v, w computed using the PINN. The…
Figure 15
Figure 15. Figure 15: Temperature field T and pressure field p in the diagonal plane of the cubic cell containing the LSC constructed with the PINN based on experimental PTV tracks [PITH_FULL_IMAGE:figures/full_fig_p020_15.png]
Figure 16
Figure 16. Figure 16: shows the probability density function (PDF) of the constructed velocity components, temperature, pressure, and heat flux w T using the PINN. In the top row of figure 16 the PDFs of the velocity components are compared with those obtained from the measured velocities.…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 11 canonical work pages

  1. [1]

    Mart \' n Abadi, Ashish Agarwal, Paul Barham, Eugene Brevdo, Zhifeng Chen, Craig Citro, Greg S. Corrado, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Ian Goodfellow, Andrew Harp, Geoffrey Irving, Michael Isard, Yangqing Jia, Rafal Jozefowicz, Lukasz Kaiser, Manjunath Kudlur, Josh Levenberg, Dan Mane, Rajat Monga, Sherry Moore, Derek Murray, ...

    arXiv doi:10.48550/arxiv.1603.04467 2016

  2. [2]

    Large-scale reorientation in cubic rayleigh--b \'e nard convection measured with particle tracking velocimetry

    Robin Barta and Claus Wagner. Large-scale reorientation in cubic rayleigh--b \'e nard convection measured with particle tracking velocimetry. Journal of Turbulence, pages 1--14, 2025. URL https://doi.org/10.1080/14685248.2024.2436843

    arXiv 2025

  3. [3]

    proptv: A probability-based particle tracking velocimetry framework

    Robin Barta, Christian Bauer, Sebastian Herzog, Daniel Schiepel, and Claus Wagner. proptv: A probability-based particle tracking velocimetry framework. Journal of computational physics, 2023. URL (under review)

    work page 2023

  4. [4]

    Physics-informed neural networks (pinns) for fluid mechanics: A review

    Shengze Cai, Zhiping Mao, Zhicheng Wang, Minglang Yin, and George Em Karniadakis. Physics-informed neural networks (pinns) for fluid mechanics: A review. Acta Mechanica Sinica, 37 0 (12): 0 1727--1738, 2021 a . URL https://doi.org/10.1007/s10409-021-01148-1

    doi:10.1007/s10409-021-01148-1 2021

  5. [5]

    Flow over an espresso cup: inferring 3-d velocity and pressure fields from tomographic background oriented S chlieren via physics-informed neural networks

    Shengze Cai, Zhicheng Wang, Frederik Fuest, Young Jin Jeon, Callum Gray, and George Em Karniadakis. Flow over an espresso cup: inferring 3-d velocity and pressure fields from tomographic background oriented S chlieren via physics-informed neural networks. Journal of Fluid Mechanics, 915, 2021 b . URL https://doi.org/10.1017/jfm.2021.135

    doi:10.1017/jfm.2021.135 2021

  6. [6]

    Scaling of hard thermal turbulence in rayleigh-b \'e nard convection

    Bernard Castaing, Gemunu Gunaratne, Francois Heslot, and Gianluigi Zanetti. Scaling of hard thermal turbulence in rayleigh-b \'e nard convection. Journal of Fluid Mechanics, 204: 0 1--30, 1989. URL https://doi.org/10.1017/S0022112089001643

    doi:10.1017/s0022112089001643 1989

  7. [7]

    Keras, 2015

    Fran c ois Chollet et al. Keras, 2015. URL https://keras.io

    2015

  8. [8]

    Reconstructing rayleigh--b \'e nard flows out of temperature-only measurements using physics-informed neural networks

    Patricio Clark Di Leoni, Lokahith Agasthya, Michele Buzzicotti, and Luca Biferale. Reconstructing rayleigh--b \'e nard flows out of temperature-only measurements using physics-informed neural networks. The European Physical Journal E, 46 0 (3): 0 16, 2023. URL https://doi.org/10.1140/epje/s10189-023-00276-9

    doi:10.1140/epje/s10189-023-00276-9 2023

Show all 36 references
  1. [9]

    The validity of the boussinesq approximation for liquids and gases

    Donald D Gray and Aldo Giorgini. The validity of the boussinesq approximation for liquids and gases. International Journal of Heat and Mass Transfer, 19 0 (5): 0 545--551, 1976. URL https://doi.org/10.1016/0017-9310(76)90168-X

    1976 doi

  2. [10]

    Scaling in thermal convection: A unifying theory

    Siegfried Grossmann and Detlef Lohse. Scaling in thermal convection: A unifying theory. Journal of Fluid Mechanics, 407: 0 27--56, 2000. URL https://doi.org/10.1017/S0022112099007545

    doi

  3. [11]

    Thermal convection for large prandtl numbers

    Siegfried Grossmann and Detlef Lohse. Thermal convection for large prandtl numbers. Physical Review Letters, 86: 0 3316--3319, 2001. URL https://doi.org/10.1103/PhysRevLett.86.3316

    2001 doi

  4. [12]

    Prandtl and Rayleigh number dependence of the Reynolds number in turbulent thermal convection

    Siegfried Grossmann and Detlef Lohse. Prandtl and Rayleigh number dependence of the Reynolds number in turbulent thermal convection. Physical Review E, 66, 2002. URL https://doi.org/10.1103/PhysRevE.66.016305

    2002 doi

  5. [13]

    Fluctuations in turbulent Rayleigh - Bénard convection: The role of plumes

    Siegfried Grossmann and Detlef Lohse. Fluctuations in turbulent Rayleigh - Bénard convection: The role of plumes. Physics of Fluids, 16: 0 4462--4472, 2004. URL https://doi.org/10.1063/1.1807751

    2004 doi

  6. [14]

    u nther Gr \

    G \"u nther Gr \"o tzbach. Spatial resolution requirements for direct numerical simulation of the Rayleigh - Bénard convection. Journal of Computational Physics, 49: 0 241--264, 1983. URL https://doi.org/10.1016/0021-9991(83)90125-0

    1983 doi

  7. [15]

    A probabilistic particle tracking framework for guided and brownian motion systems with high particle densities

    Sebastian Herzog, Daniel Schiepel, Isabella Guido, Robin Barta, and Claus Wagner. A probabilistic particle tracking framework for guided and brownian motion systems with high particle densities. SN Computer Science, 2: 0 1--20, 2021. URL https://doi.org/10.1007/s42979-021-00879-z

    2021 doi

  8. [16]

    Enhancing pinns for solving pdes via adaptive collocation point movement and adaptive loss weighting

    Jie Hou, Ying Li, and Shihui Ying. Enhancing pinns for solving pdes via adaptive collocation point movement and adaptive loss weighting. Nonlinear Dynamics, 111 0 (16): 0 15233--15261, 2023. URL https://doi.org/10.1007/s11071-023-08654-w

    2023 doi

  9. [17]

    Volumetric lagrangian temperature and velocity measurements with thermochromic liquid crystals

    T K \"a ufer and C Cierpka. Volumetric lagrangian temperature and velocity measurements with thermochromic liquid crystals. Measurement Science and Technology, 35 0 (3): 0 035301, 2023. URL https://doi.org/10.1088/1361-6501/ad16d1

    2023 doi

  10. [18]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint, 2014. URL https://doi.org/10.48550/ARXIV.1412.6980

    2014 doi arXiv

  11. [19]

    Simple computational strategies for more effective physics-informed neural networks modeling of turbulent natural convection

    Didier Lucor, Atul Agrawal, and Anne Sergent. Simple computational strategies for more effective physics-informed neural networks modeling of turbulent natural convection. Journal of Computational Physics, 456: 0 111022, 2022. URL https://doi.org/10.1016/j.jcp.2022.111022

    2022

  12. [20]

    Particle tracking velocimetry in three-dimensional flows

    HG Maas, A Gruen, and D Papantoniou. Particle tracking velocimetry in three-dimensional flows. Experiments in Fluids, 15: 0 133--146, 1993. URL https://doi.org/10.1007/BF00190953

    1993 doi

  13. [21]

    Particle tracking velocimetry in three-dimensional flows

    NA Malik, Th Dracos, and DA Papantoniou. Particle tracking velocimetry in three-dimensional flows. Experiments in Fluids, 15: 0 279--294, 1993. URL https://doi.org/10.1007/BF00223406

    1993 doi

  14. [22]

    Periodically activated physics-informed neural networks for assimilation tasks for three-dimensional rayleigh–bénard convection

    Michael Mommert, Robin Barta, Christian Bauer, Marie-Christine Volk, and Claus Wagner. Periodically activated physics-informed neural networks for assimilation tasks for three-dimensional rayleigh–bénard convection. Computers and Fluids, 283: 0 106419, 2024. URL https://doi.or...

    2024

  15. [23]

    Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations

    Maziar Raissi, Paris Perdikaris, and George E Karniadakis. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational physics, 378: 0 686--707, 2019. URL http...

    2019 doi

  16. [24]

    On convection currents in a horizontal layer of fluid, when the higher temperature is on the under side

    Lord Rayleigh. On convection currents in a horizontal layer of fluid, when the higher temperature is on the under side. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science, 32 0 (192): 0 529–546, 1916. URL https://doi.org/10.1080/14786441608635602

    1916 doi

  17. [25]

    On the impact of larger batch size in the training of physics informed neural networks

    Shyam Sankaran, Hanwen Wang, Leonardo Ferreira Guilhoto, and Paris Perdikaris. On the impact of larger batch size in the training of physics informed neural networks. In The Symbiosis of Deep Learning and Differential Equations II, 2022. URL https://openreview.net/forum?id=THCvohg1RV

    2022

  18. [26]

    Resolving the fine-scale structure in turbulent rayleigh--b \'e nard convection

    Janet D Scheel, Mohammad S Emran, and J \"o rg Schumacher. Resolving the fine-scale structure in turbulent rayleigh--b \'e nard convection. New Journal of Physics, 15 0 (11): 0 113063, 2013. URL https://doi.org/10.1088/1367-2630/15/11/113063

    2013 doi

  19. [27]

    Numerical and experimental study of aerosol dispersion in the D o728 aircraft cabin

    Daniel Schmeling, Andrei Shishkin, Daniel Schiepel, and Claus Wagner. Numerical and experimental study of aerosol dispersion in the D o728 aircraft cabin. CEAS Aeronautical Journal, 14: 0 509526, 2023. https://doi.org/10.1007/s13272-023-00644-3

    2023 doi

  20. [28]

    3d Lagrangian particle tracking in fluid mechanics

    Andreas Schr \"o der and Daniel Schanz. 3d Lagrangian particle tracking in fluid mechanics. Annual Review of Fluid Mechanics, 55, 2023. URL https://doi.org/10.1146/annurev-fluid-031822-041721

    2023 doi

  21. [29]

    Boundary layer structure in turbulent thermal convection and its consequences for the required numerical resolution

    Olga Shishkina, Richard J A M Stevens, Siegfried Grossmann, and Detlef Lohse. Boundary layer structure in turbulent thermal convection and its consequences for the required numerical resolution. New Journal of Physics, 12, 2010. URL https://doi.org/10.1088/1367-2630/12/7/075022

    2010 doi

  22. [30]

    Vincent Sitzmann, Julien N. P. Martel, Alexander W. Bergman, David B. Lindell, and Gordon Wetzstein. Implicit neural representations with periodic activation functions. arXiv preprint, 2020. URL https://doi.org/10.48550/ARXIV.2006.09661

    2020 doi arXiv

  23. [31]

    Distortion compensation for generalized stereoscopic particle image velocimetry

    Steven M Soloff, Ronald J Adrian, and Zi-Chao Liu. Distortion compensation for generalized stereoscopic particle image velocimetry. Measurement Science and Technology, 8, 1997. URL https://doi.org/10.1088/0957-0233/8/12/008

    1997 doi

  24. [32]

    Inferring turbulent velocity and temperature fields and their statistics from lagrangian velocity measurements using physics-informed kolmogorov-arnold networks

    Juan Diego Toscano, Theo Käufer, Zhibo Wang, Martin Maxey, Christian Cierpka, and George Em Karniadakis. Inferring turbulent velocity and temperature fields and their statistics from lagrangian velocity measurements using physics-informed kolmogorov-arnold networks. arXiv prep...

    2024 doi arXiv

  25. [33]

    A pinn methodology for temperature field reconstruction in the piv measurement plane: Case of rayleigh-b 'enard convection

    Marie-Christine Volk, Anne Sergent, Didier Lucor, Michael Mommert, Christian Bauer, and Claus Wagner. A pinn methodology for temperature field reconstruction in the piv measurement plane: Case of rayleigh-b 'enard convection. arXiv preprint arXiv:2503.23801, 2025

    2025 arXiv

  26. [34]

    Comments on the numerical investigation of Rayleigh and Marangoni convection in a vertical circular cylinder

    Claus Wagner, Rainer Friedrich, and Ranga Narayanan. Comments on the numerical investigation of Rayleigh and Marangoni convection in a vertical circular cylinder. Physics of Fluids, 6: 0 1425--1433, 1994. URL https://doi.org/10.1063/1.868257

    1994 doi

  27. [35]

    What rayleigh numbers are achievable under oberbeck–boussinesq conditions? Journal of Fluid Mechanics, 986 0 (R2), 2024

    S Weiss, Ms Emran, and O Shishkina. What rayleigh numbers are achievable under oberbeck–boussinesq conditions? Journal of Fluid Mechanics, 986 0 (R2), 2024. URL https://doi.org/10.1017/jfm.2024.389

    2024 doi

  28. [36]

    Volume self-calibration for 3d particle image velocimetry

    B Wieneke. Volume self-calibration for 3d particle image velocimetry. Experiments in Fluids, 45: 0 549--556, 2008. URL https://doi.org/10.1007/s00348-008-0521-5

    2008 doi

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.