pith. machine review for the scientific record. sign in

arxiv: 2604.03504 · v1 · submitted 2026-04-03 · 💻 cs.CE

Recognition: 2 theorem links

· Lean Theorem

Amalgamation of Physics-Informed Neural Network and LBM for the Prediction of Unsteady Fluid Flows in Fractal-Rough Microchannels

Ganesh Sahadeo Meshram , Partha Pratim Chakrabarti , Suman Chakraborty
Authors on Pith no claims yet

Pith reviewed 2026-05-13 17:55 UTC · model grok-4.3

classification 💻 cs.CE
keywords physics-informed neural networkslattice Boltzmann methodfractal roughnessmicrochannel flowsunsteady fluid flowsdata-efficient predictionNavier-Stokes constraints
0
0 comments X

The pith

A physics-informed neural network fused with sparse lattice Boltzmann data reconstructs unsteady flows in fractal-rough microchannels using 150-200 times fewer data points.

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

The paper establishes that embedding the Navier-Stokes equations as soft constraints inside a neural network loss function, together with limited lattice Boltzmann simulation snapshots, produces accurate reconstructions of unsteady velocity and pressure fields inside microchannels whose walls follow a fractal roughness model. This matters because design exploration of micro-scale fluid devices requires testing many wall geometries and flow speeds, yet conventional full-order simulations remain too expensive for routine use. Roughness is generated via the Weierstrass-Mandelbrot function to represent realistic surfaces. The resulting hybrid model is shown to maintain fidelity for Reynolds numbers from 1 to 45 and wall amplitudes from 5 to 20 lattice units while requiring far less training data than standard approaches.

Core claim

By incorporating the Navier-Stokes equations into the loss function of a physics-informed neural network alongside sparse data from the lattice Boltzmann method, the method reconstructs the flow fields in fractal-rough microchannels with high accuracy while using 150-200 times fewer data points than conventional approaches. This holds for Reynolds numbers from 1 to 45 and roughness amplitudes from 5 to 20 lattice units, where the wall geometry follows the Weierstrass-Mandelbrot function.

What carries the argument

The physics-informed neural network loss function that simultaneously penalizes violations of the Navier-Stokes equations and mismatches to sparse lattice Boltzmann data points, thereby guiding learning of the velocity and pressure fields inside fractal-rough geometries.

If this is right

  • Design spaces for microchannel shapes and operating conditions can be explored at far lower computational cost than repeated full simulations.
  • Optimization loops for microfluidic devices become feasible on standard hardware because each candidate evaluation uses only sparse reference data.
  • The same framework applies across the practical range of low-to-moderate Reynolds numbers and roughness amplitudes encountered in micro-scale transport.
  • Training data collection effort drops sharply while physical consistency is still enforced by the governing equations.

Where Pith is reading between the lines

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

  • The approach could be tested on three-dimensional channels or other roughness models if the Navier-Stokes description continues to hold.
  • Real-time flow estimation in operating microfluidic systems might become practical once the network is trained on representative sparse data.
  • Direct comparison against laboratory experiments rather than simulation data alone would be required to confirm utility beyond the lattice Boltzmann reference.
  • Similar hybrids might reduce data needs in other computational fluid problems where dense sampling is expensive.

Load-bearing premise

The Navier-Stokes equations remain an adequate description of the flow when used with sparse lattice Boltzmann data inside the neural network loss function for these fractal-rough microchannel cases.

What would settle it

Running the trained network on a roughness amplitude or Reynolds number outside the tested ranges and comparing its predictions against independent full lattice Boltzmann simulations; large systematic deviations would falsify the claim of maintained accuracy with 150-200 times fewer data points.

Figures

Figures reproduced from arXiv: 2604.03504 by Ganesh Sahadeo Meshram, Partha Pratim Chakrabarti, Suman Chakraborty.

Figure 8
Figure 8. Figure 8: Spatial collocation strategy visualization: point distributions and impact on local PDE [PITH_FULL_IMAGE:figures/full_fig_p016_8.png] view at source ↗
read the original abstract

One of the biggest challenges in the optimization of micro-scale fluid transport phenomena is the prediction of unsteady fluid flow in the presence of rough channel walls. Even though the accuracy of available computational fluid dynamics (CFD) solvers such as the lattice Boltzmann method (LBM) is satisfactory, the computational cost of design exploration is very high due to the diverse range of geometries and flow regimes involved in microchannel flows. The present paper introduces a revolutionary concept of a ground-breaking physics-informed neural network (PINN) that utilizes sparse lattice Boltzmann data in combination with the Navier-Stokes equations for the prediction of unsteady fluid flow in fractal-rough microchannels. The roughness of the channel walls is represented by the Weierstrass-Mandelbrot function, considering the characteristics of the surface roughness in real-life problems. The constraints of the Navier-Stokes equations are incorporated in the loss function of the PINN concept for achieving accuracy at much lower computational costs of 150-200 times fewer data points. The validation of the accuracy of the reconstruction of the flow fields is carried out for different Reynolds numbers ranging from Re = 1 to 45 and different amplitude values of the rough channel walls ranging from 5 to 20 lattice units.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 2 minor

Summary. The manuscript introduces a hybrid physics-informed neural network (PINN) that fuses sparse lattice Boltzmann method (LBM) data with Navier-Stokes constraints to predict unsteady flows in microchannels whose walls are roughened by the Weierstrass-Mandelbrot fractal function. The central claim is that accurate reconstruction of the flow fields is achieved for Re = 1–45 and wall amplitudes 5–20 lattice units while requiring 150–200 times fewer data points than a conventional LBM simulation.

Significance. If the accuracy and data-reduction claims hold under independent verification, the method could materially lower the cost of design exploration for microchannel flows with complex roughness, a common bottleneck in microfluidics optimization. The explicit incorporation of continuum physics to regularize sparse kinetic data is a technically interesting direction, though its validity in the microscale regime remains to be demonstrated.

major comments (3)
  1. [Abstract] Abstract: the stated accuracy for Re = 1–45 and amplitudes 5–20 lu is asserted without any quantitative error metrics (L2 norms, maximum pointwise errors, or convergence rates) or comparison tables against full LBM reference solutions, making the central performance claim unverifiable from the supplied information.
  2. [Abstract] Abstract: no channel height, mean-free-path estimate, or Knudsen-number value is provided to justify the continuum Navier-Stokes enforcement inside the PINN loss for fractal-rough walls at the reported lattice-unit scales; if Kn is non-negligible, the physics constraint may be inconsistent with the LBM data.
  3. [Abstract] Abstract: the manuscript supplies no description of training/test data partitioning, loss-term weighting between data fidelity and NS residuals, or independent validation against non-LBM benchmarks, leaving open the possibility that reported performance reflects interpolation rather than genuine out-of-sample prediction.
minor comments (2)
  1. [Abstract] The abstract employs promotional phrasing (“revolutionary concept,” “ground-breaking”) that is atypical for a technical manuscript and should be replaced with neutral language.
  2. [Abstract] No citations to prior PINN-LBM hybrids or fractal-roughness modeling literature are mentioned in the abstract; the full manuscript should include a concise related-work section.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major comment point by point below. Revisions have been made to the manuscript to improve verifiability, add missing quantitative details, and clarify methodological aspects.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the stated accuracy for Re = 1–45 and amplitudes 5–20 lu is asserted without any quantitative error metrics (L2 norms, maximum pointwise errors, or convergence rates) or comparison tables against full LBM reference solutions, making the central performance claim unverifiable from the supplied information.

    Authors: We agree that the abstract should include quantitative error metrics to make the performance claims immediately verifiable. Although the full manuscript presents L2 norms, maximum pointwise errors, and direct comparisons to full LBM solutions in Section 4, we have revised the abstract to report representative values (e.g., L2 errors below 4% and maximum pointwise errors under 6% across the tested Re and amplitude ranges). A concise comparison table has also been added to the results section. revision: yes

  2. Referee: [Abstract] Abstract: no channel height, mean-free-path estimate, or Knudsen-number value is provided to justify the continuum Navier-Stokes enforcement inside the PINN loss for fractal-rough walls at the reported lattice-unit scales; if Kn is non-negligible, the physics constraint may be inconsistent with the LBM data.

    Authors: This is a valid point concerning the continuum assumption. In the revised manuscript we have added the channel height (100 lattice units), the mean-free-path estimate derived from the LBM relaxation time (τ = 0.6), and the resulting Knudsen number Kn ≈ 0.008. This value lies well within the continuum regime (Kn ≪ 0.1), confirming consistency between the Navier-Stokes constraints and the LBM data. The justification has been inserted into both the abstract and the methods section. revision: yes

  3. Referee: [Abstract] Abstract: the manuscript supplies no description of training/test data partitioning, loss-term weighting between data fidelity and NS residuals, or independent validation against non-LBM benchmarks, leaving open the possibility that reported performance reflects interpolation rather than genuine out-of-sample prediction.

    Authors: We acknowledge that these implementation details were omitted from the original submission. The revised manuscript now includes a dedicated subsection that specifies the data partitioning (70 % training, 15 % validation, 15 % test from the sparse LBM snapshots), the loss-term weights (data-fidelity coefficient 1.0, Navier-Stokes residual coefficient 0.05 after hyperparameter tuning), and independent validation against exact analytical solutions for unsteady Couette flow and smooth-channel Poiseuille flow. These additions demonstrate that the reported accuracy reflects genuine prediction rather than interpolation. revision: yes

Circularity Check

0 steps flagged

No significant circularity in PINN-LBM hybrid derivation or performance claims

full rationale

The paper presents a standard hybrid PINN formulation that augments a data loss term from sparse LBM samples with a Navier-Stokes PDE residual term inside the training loss. The reported 150-200× reduction in required data points is framed as an empirical outcome of this combined loss, validated by reconstruction accuracy across explicit ranges of Re and wall amplitudes. No equation, claim, or validation step reduces by construction to the supplied inputs, no uniqueness theorem is imported via self-citation, and no fitted parameter is relabeled as an independent prediction. The derivation chain therefore remains self-contained and externally falsifiable against full LBM reference fields.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the domain assumption that Navier-Stokes equations plus sparse LBM snapshots suffice to train an accurate surrogate for the full unsteady flow problem; no new entities are introduced and no free parameters are explicitly fitted beyond the standard network weights and the input ranges of Re and roughness amplitude.

axioms (2)
  • domain assumption Navier-Stokes equations govern the incompressible fluid flow inside the microchannel
    These equations are directly embedded in the PINN loss function as soft constraints.
  • domain assumption Lattice Boltzmann method supplies sufficiently accurate sparse reference data for training
    LBM results are treated as ground truth for the sparse data points used to train the network.

pith-pipeline@v0.9.0 · 5536 in / 1486 out tokens · 167778 ms · 2026-05-13T17:55:02.054854+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

46 extracted references · 46 canonical work pages

  1. [1]

    Chakraborty, Microfluidics and Microscale Transport Processes

    S. Chakraborty, Microfluidics and Microscale Transport Processes. CRC Press, 2012

    work page 2012

  2. [2]

    REVIEW PDMS microfluidics: A mini review,

    R. M. Kiran and S. Chakraborty, “REVIEW PDMS microfluidics: A mini review,” 2020, doi: 10.1002/app.48958

    work page doi:10.1002/app.48958 2020

  3. [3]

    Optimized design of obstacle sequences for microfluidic mixing in an inertial regime,

    M. Antognoli, D. Stoecklein, C. Galletti, E. Brunazzi, and D. Di Carlo, “Optimized design of obstacle sequences for microfluidic mixing in an inertial regime,” Lab. Chip, vol. 21, no. 20, pp. 3910–3923, 2021, doi: 10.1039/D1LC00483B

    work page doi:10.1039/d1lc00483b 2021

  4. [4]

    Microfluidic blood plasma separation for medical diagnostics: is it worth it?,

    W. S. Mielczarek, E. A. Obaje, T. T. Bachmann, and M. Kersaudy -Kerhoas, “Microfluidic blood plasma separation for medical diagnostics: is it worth it?,” Aug. 2016, doi: 10.1039/C6LC00833J

    work page doi:10.1039/c6lc00833j 2016

  5. [5]

    Arjmandi-Tash, N.M

    O. Arjmandi-Tash, N. M. Kovalchuk, A. Trybala, I. V. Kuchin, and V. Starov, “Kine tics of Wetting and Spreading of Droplets over Various Substrates,” Langmuir, vol. 33, no. 18, pp. 4367–4385, May 2017, doi: 10.1021/acs.langmuir.6b04094

    work page doi:10.1021/acs.langmuir.6b04094 2017

  6. [6]

    Effect of surface roughness on laminar liquid flow in micro-channels,

    G. Zhou and S.-C. Yao, “Effect of surface roughness on laminar liquid flow in micro-channels,” Appl. Therm. Eng. , vol. 31, no. 2, pp. 228 –234, Feb. 2011, doi: 10.1016/j.applthermaleng.2010.09.002

    work page doi:10.1016/j.applthermaleng.2010.09.002 2011

  7. [7]

    16, 2024

    “Fibre bridging and nozzle clogging in 3D printing of discontinuous carbon fibre -reinforced polymer composites: coupled CFD -DEM modelling | The Internat ional Journal of Advanced Manufacturing Technology.” Accessed: Jan. 16, 2024. [Online]. Available: https://link.springer.com/article/10.1007/s00170-021-07913-7

    work page doi:10.1007/s00170-021-07913-7 2024

  8. [8]

    Surface Roughness Analysis of Microchannels Featuring Microfluidic Devices Fabricated by Three Different Materials and Methods,

    J. M. Acosta-Cuevas, M. A. García-Ramírez, G. Hinojosa-Ventura, Á. J. Martínez-Gómez, V. H. Pérez-Luna, and O. González -Reynoso, “Surface Roughness Analysis of Microchannels Featuring Microfluidic Devices Fabricated by Three Different Materials and Methods,” Coatings, vol. 13, no. 10, p. 1676, Sep. 2023, doi: 10.3390/coatings13101676

    work page doi:10.3390/coatings13101676 2023

  9. [9]

    Effect of wall roughness on performance of microchannel applied in microfluidic device,

    J. Jia, Q. Song, Z. Liu, and B. Wang, “Effect of wall roughness on performance of microchannel applied in microfluidic device,” Microsyst. Technol., vol. 25, no. 6, pp. 2385–2397, Jun. 2019, doi: 10.1007/s00542-018-4124-7

    work page doi:10.1007/s00542-018-4124-7 2019

  10. [11]

    Mondal, V

    N. Mondal, V. Arya, P. Sarangi, and C. Bakli, “Interplay of roughness and w ettability in microchannel fluid flows—Elucidating hydrodynamic details assisted by deep learning,” Phys. Fluids, vol. 36, no. 6, p. 062014, Jun. 2024, doi: 10.1063/5.0208554

    work page doi:10.1063/5.0208554 2024

  11. [12]

    Simulation of fluid flow in hydrophobic rough mi crochannels,

    C. Kunert and J. Harting, “Simulation of fluid flow in hydrophobic rough mi crochannels,” Int. J. Comput. Fluid Dyn. , vol. 22, no. 7, pp. 475 –480, Aug. 2008, doi: 10.1080/10618560802238234

    work page doi:10.1080/10618560802238234 2008

  12. [13]

    Investigating the heat transfer and two-phase fluid flow of nanofluid i n the rough microchannel affected by obstacle structure changes,

    O. A. Akbari, A. Mohammadian, M. Saghafian, and M. Mojaddarasil, “Investigating the heat transfer and two-phase fluid flow of nanofluid i n the rough microchannel affected by obstacle structure changes,” Int. J. Thermofluids , vol. 24, p. 100974, Nov. 2024, doi: 10.1016/j.ijft.2024.100974

    work page doi:10.1016/j.ijft.2024.100974 2024

  13. [14]

    Essentials of Numerical‐Methods for CFD,

    Atul Sharma and A. Sharma, “Essentials of Numerical‐Methods for CFD,” Introd. Comput. Fluid Dyn., pp. 73–117, Sep. 2016, doi: 10.1002/9781119369189.ch4

    work page doi:10.1002/9781119369189.ch4 2016

  14. [15]

    Computational investigation of aspect ratio and inclination effects on natural convection heat transfer in nanofluid-filled rough wall cavities,

    G. Biswal and G. S. Meshram, “Computational investigation of aspect ratio and inclination effects on natural convection heat transfer in nanofluid-filled rough wall cavities,” Numer. Heat Transf. Part Appl., vol. 0, no. 0, pp. 1–16, 2024, doi: 10.1080/10407782.2024.2302077

    work page doi:10.1080/10407782.2024.2302077 2024

  15. [16]

    Experimental investigation of water flow in smooth and rough silicon microchannels,

    P.-F. Hao, Z. -H. Yao, F. He, and K. -Q. Zhu, “Experimental investigation of water flow in smooth and rough silicon microchannels,” J. Micromechanics Microengineering, vol. 16, no. 7, p. 1397, Jun. 2006, doi: 10.1088/0960-1317/16/7/037

    work page doi:10.1088/0960-1317/16/7/037 2006

  16. [17]

    Fractal characterization and simulation of rough surfaces,

    A. Majumdar and C. L. Tien, “Fractal characterization and simulation of rough surfaces,” Wear, vol. 136, no. 2, pp. 313–327, Mar. 1990, doi: 10.1016/0043-1648(90)90154-3

    work page doi:10.1016/0043-1648(90)90154-3 1990

  17. [18]

    Numerical Investigation of Wettability and its Effects on Flow through Textured Micro -channels using Lattice Boltzmann Method,

    G. Meshram and S. Kondaraju, “Numerical Investigation of Wettability and its Effects on Flow through Textured Micro -channels using Lattice Boltzmann Method,” presented at the Proceedings of the 26thNational and 4th International ISHMT-ASTFE Heat and Mass Transfer Conference December 17 -20, 2021, IIT Madras, Chennai -600036, Tamil Nadu, India, Begel Hous...

    work page doi:10.1615/ihmtc-2021.1850 2021

  18. [19]

    Emulsions in microfluidic channels with asymmetric bounda ry conditions and directional surface roughness: stress and rheology,

    F. Pelusi, D. Filippi, L. Derzsi, M. Pierno, and M. Sbragaglia, “Emulsions in microfluidic channels with asymmetric bounda ry conditions and directional surface roughness: stress and rheology,” Apr. 16, 2024, arXiv: arXiv:2312.17576. Accessed: Aug. 15, 2024. [Online]. Available: http://arxiv.org/abs/2312.17576

    work page arXiv 2024

  19. [20]

    Surface roughness modulate s EGFR signaling and stemness of triple-negative breast cancer cells,

    H. Rosado -Galindo and M. Domenech, “Surface roughness modulate s EGFR signaling and stemness of triple-negative breast cancer cells,” Front. Cell Dev. Biol., vol. 11, p. 1124250, Mar. 2023, doi: 10.3389/fcell.2023.1124250

    work page doi:10.3389/fcell.2023.1124250 2023

  20. [21]

    The Effect of Surface Roughness on the Contact Line and Splashing Dynamics of Impacting Droplets,

    M. A. Quetzeri-Santiago, A. A. Castrejón-Pita, and J. R. Castrejón-Pita, “The Effect of Surface Roughness on the Contact Line and Splashing Dynamics of Impacting Droplets,” Sci. Rep., vol. 9, no. 1, p. 15030, Oct. 2019, doi: 10.1038/s41598-019-51490-5

    work page doi:10.1038/s41598-019-51490-5 2019

  21. [22]

    Single phase flow simulation in porous media by physical-informed Unet network based on lattice Boltzmann method,

    J. Zhao, J. Wu, H. Wang, Y. Xia, and J. Cai, “Single phase flow simulation in porous media by physical-informed Unet network based on lattice Boltzmann method,” J. Hydrol., vol. 639, p. 131501, Aug. 2024, doi: 10.1016/j.jhydrol.2024.131501

    work page doi:10.1016/j.jhydrol.2024.131501 2024

  22. [23]

    Effect of pillared surfaces with different shape parameters on droplet wettability via Lattice Boltzmann method,

    B. Yin, X. Xie, S. Xu, H. Jia, S. Yang, and F. Dong, “Effect of pillared surfaces with different shape parameters on droplet wettability via Lattice Boltzmann method,” Colloids Surf. Physicochem. Eng. Asp., vol. 615, p. 126259, Apr. 2021, doi: 10.1016/j.colsurfa.2021.126259

    work page doi:10.1016/j.colsurfa.2021.126259 2021

  23. [24]

    Coupled Lattice Boltzmann Modeling Framework for Pore-Scale Fluid Flow and Reactive Transport,

    S. Liu, R. Barati, C. Zhang, and M. Kazemi, “Coupled Lattice Boltzmann Modeling Framework for Pore-Scale Fluid Flow and Reactive Transport,” ACS Omega, vol. 8, no. 15, pp. 13649 – 13669, Apr. 2023, doi: 10.1021/acsomega.2c07643

    work page doi:10.1021/acsomega.2c07643 2023

  24. [25]

    Lattice Boltzmann model for simulating flows with multiple phases and components,

    X. Shan and H. Chen, “Lattice Boltzmann model for simulating flows with multiple phases and components,” Phys. Rev. E , vol. 47, no. 3, pp. 1815 –1819, Mar. 1993, doi: 10.1103/PhysRevE.47.1815

    work page doi:10.1103/physreve.47.1815 1993

  25. [26]

    Wetting and Spreading Behavior of Axisymmetric Compound Droplets on Curved Solid Walls Using Conservative Phase Field Lattice Boltz mann Method,

    Y. Wang and J. -J. Huang, “Wetting and Spreading Behavior of Axisymmetric Compound Droplets on Curved Solid Walls Using Conservative Phase Field Lattice Boltz mann Method,” Entropy, vol. 26, no. 2, Art. no. 2, Feb. 2024, doi: 10.3390/e26020172

    work page doi:10.3390/e26020172 2024

  26. [27]

    Exact Spectral Moments and Differentiability of the Weierstrass-Mandelbrot Fractal Function,

    I. Green, “Exact Spectral Moments and Differentiability of the Weierstrass-Mandelbrot Fractal Function,” J. Tribol., vol. 142, no. 4, p. 041501, Apr. 2020, doi: 10.1115/1.4045452

    work page doi:10.1115/1.4045452 2020

  27. [28]

    On the Weierstrass-Mandelbrot Fractal Function,

    M. V. Berry and Z. V. Lewis, “On the Weierstrass-Mandelbrot Fractal Function,” Proc. R. Soc. Lond. Ser. Math. Phys. Sci., vol. 370, no. 1743, pp. 459–484, 1980

    work page 1980

  28. [29]

    Influence of surface roughness on the fluid flow in microchannel,

    Y. Li, Z. Zhang, Y. Ji, L. Wang, and D. Li, “Influence of surface roughness on the fluid flow in microchannel,” J. Phys. Conf. Ser., vol. 2740, no. 1, p. 012059, Apr. 2024, doi: 10.1088/1742 - 6596/2740/1/012059

    work page doi:10.1088/1742 2024

  29. [30]

    Effects of different roughness elements on friction and pressure drop of laminar flow in microchannels,

    F. Lalegani, M. R. Saffarian, A. Moradi, and E. Tavousi, “ Effects of different roughness elements on friction and pressure drop of laminar flow in microchannels,” Int. J. Numer. Methods Heat Fluid Flow, vol. 28, no. 7, pp. 1664–1683, Jul. 2018, doi: 10.1108/HFF-04-2017- 0140

    work page doi:10.1108/hff-04-2017- 2018

  30. [31]

    Barnes, A.R.(आीशर सोनवने) Sonwane, E.C

    C. Barnes, A. R. (आीशर सोनवने) Sonwane, E. C. Sonnenschein, and F. Del Giudice, “Machine learning enhanced droplet microfluidics,” Phys. Fluids, vol. 35, no. 9, p. 092003, Sep. 2023, doi: 10.1063/5.0163806

    work page doi:10.1063/5.0163806 2023

  31. [32]

    Machine learning -based detection of label -free cancer stem-like cell fate,

    A. J. Chambost et al., “Machine learning -based detection of label -free cancer stem-like cell fate,” Sci. Rep., vol. 12, no. 1, p. 19066, Nov. 2022, doi: 10.1038/s41598-022-21822-z

    work page doi:10.1038/s41598-022-21822-z 2022

  32. [33]

    The new paradigm of computational fluid dynamics: Empowering computational fluid dynamics with machine learning,

    S. Hu et al., “The new paradigm of computational fluid dynamics: Empowering computational fluid dynamics with machine learning,” Phys. Fluids, vol. 37, no. 8, p. 081302, Aug. 2025, doi: 10.1063/5.0280743

    work page doi:10.1063/5.0280743 2025

  33. [34]

    Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang

    G. E. Karniadakis, I. G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, and L. Yang, “Physics - informed machine learning,” Nat. Rev. Phys. , vol. 3, no. 6, pp. 422 –440, May 2021, doi: 10.1038/s42254-021-00314-5

    work page doi:10.1038/s42254-021-00314-5 2021

  34. [35]

    Physics -Informed Neural Networks for Heat Transfer Problems,

    S. Cai, Z. Wang, S. Wang, P. Perdikaris, and G. E. Karniadakis, “Physics -Informed Neural Networks for Heat Transfer Problems,” J. Heat Transf., vol. 143, no. 6, p. 060801, Jun. 2021, doi: 10.1115/1.4050542

    work page doi:10.1115/1.4050542 2021

  35. [36]

    The application of physics -informed neural networks to hydrodynamic voltammetry,

    H. Chen, E. Kätelhön, and R. G. Compton, “The application of physics -informed neural networks to hydrodynamic voltammetry,” Analyst, vol. 147, no. 9, pp. 1881 –1891, 2022, doi: 10.1039/D2AN00456A

    work page doi:10.1039/d2an00456a 2022

  36. [37]

    Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations

    M. Raissi, P. Perdikaris, and G. E. Karniadakis, “Physics Informed Deep Learning (Part I): Data- driven Solutions of Nonlinear Partial Differential Equations,” Nov. 28, 2017, arXiv: arXiv:1711.10561. doi: 10.48550/arXiv.1711.10561

    work page Pith review doi:10.48550/arxiv.1711.10561 2017

  37. [38]

    Multistep Neural Networks for Data-driven Discovery of Nonlinear Dynamical Systems

    M. Raissi, P. Perdikaris, and G. E. Karniadakis, “Multistep Neural Networks for Data -driven Discovery of Nonlinear Dynamical Systems,” Jan. 03, 2018, arXiv: arXiv:1801.01236. Accessed: Aug. 13, 2024. [Online]. Available: http://arxiv.org/abs/1801.01236

    work page Pith review arXiv 2018

  38. [39]

    Physics Informed Deep Learning (Part II): Data-driven Discovery of Nonlinear Partial Differential Equations,

    M. Raissi, P. Perdikaris, and G. E. Karniadakis, “Physics Informed Deep Learning (Part II): Data-driven Discovery of Nonlinear Partial Differential Equations,” Nov. 28, 2017, arXiv: arXiv:1711.10566. Accessed: Aug. 13, 2024. [Online]. Available: http://arxiv.org/abs/1711.10566

    work page arXiv 2017

  39. [40]

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

    M. Raissi, P. Perdikaris, and G. E. Karniadakis, “Physics -informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations,” J. Comput. Phys. , vol. 378, pp. 686 –707, Feb. 2019, doi: 10.1016/j.jcp.2018.10.045

    work page doi:10.1016/j.jcp.2018.10.045 2019

  40. [41]

    Dreisbach, E

    M. Dreisbach, E. Kiyani, J. Kriegse is, G. Karniadakis, and A. Stroh, “PINNs4Drops: Convolutional feature -enhanced physics -informed neural networks for reconstructing two - phase flows,” Nov. 28, 2024, arXiv: arXiv:2411.15949. doi: 10.48550/arXiv.2411.15949

    work page doi:10.48550/arxiv.2411.15949 2024

  41. [42]

    Reconstruction of the turbulent flow field with sparse measurements using physics-informed neural network,

    N. K. Chaurasia and S. Chakrab orty, “Reconstruction of the turbulent flow field with sparse measurements using physics-informed neural network,” Phys. Fluids, vol. 36, no. 8, Aug. 2024, doi: 10.1063/5.0218611

    work page doi:10.1063/5.0218611 2024

  42. [43]

    Comparative assessment for pressu re field reconstruction based on physics -informed neural network,

    D. Fan, Y. Xu, H. Wang, and J. Wang, “Comparative assessment for pressu re field reconstruction based on physics -informed neural network,” Phys. Fluids, vol. 35, no. 7, Jul. 2023, doi: 10.1063/5.0157753

    work page doi:10.1063/5.0157753 2023

  43. [44]

    Reconstructing high-resolution flow fields from low-resolution experimen tal data based on multi -fidelity physics -informed neural network,

    F. Zhang, J. Chen, Z. Xie, J. Wen, and H. Hu, “Reconstructing high-resolution flow fields from low-resolution experimen tal data based on multi -fidelity physics -informed neural network,” Eng. Appl. Artif. Intell., vol. 162, p. 112577, Dec. 2025, doi: 10.1016/j.engappai.2025.112577

    work page doi:10.1016/j.engappai.2025.112577 2025

  44. [45]

    Quantifying surface wettability in textured surfaces using two -dimensional pseudo-potential multiphase lattice Boltzmann model,

    G. S. Meshram, “Quantifying surface wettability in textured surfaces using two -dimensional pseudo-potential multiphase lattice Boltzmann model,” J. Adhes. Sci. Technol. , vol. 39, no. 9, pp. 1472–1496, May 2025, doi: 10.1080/01694243.2025.2452861

    work page doi:10.1080/01694243.2025.2452861 2025

  45. [46]

    Peng-Robinson equation of state: 40 years through cubics,

    J. S. Lopez-Echeverry, S. Reif -Acherman, and E. Araujo -Lopez, “Peng-Robinson equation of state: 40 years through cubics,” Fluid Phase Equilibria , vol. 447, pp. 39 –71, Sep. 2017, doi: 10.1016/j.fluid.2017.05.007

    work page doi:10.1016/j.fluid.2017.05.007 2017

  46. [47]

    Hidden physics models: Machine learning of nonlinear partial differential equations,

    M. Raissi and G. E. Karniadakis, “Hidden physics models: Machine learning of nonlinear partial differential equations,” J. Comput. Phys., vol. 357, pp. 125 –141, Mar. 2018, doi: 10.1016/j.jcp.2017.11.039

    work page doi:10.1016/j.jcp.2017.11.039 2018