ANN INTEGRATED DLVO-SPH BASED SIMULATION FOR PREDICTING STABILITY OF HYBRID NANOFLUIDS
Advances and Applications in Fluid Mechanics · 01.08.2026 · 33 (2) · 253-278
DOI: 10.17654/0973468626012 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advances and Applications in Fluid Mechanics, online erschienen am 01.08.2026. Verfasst von Swati Kawadu Gaydhane und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The stability of hybrid nanofluids plays a crucial role in determining their thermophysical performance and long-term applicability in advanced thermal management systems. Experimental determination of nanofluid’s stability is often time-consuming, expensive, and influenced by multiple interacting physicochemical parameters. To overcome these limitations, this study investigates the colloidal stability of hybrid nanofluids using a numerical model based on the physics-informed Artificial Neural Network (ANN) integrated with Derjaguin-Landau-Verwey-Overbeek (DLVO) theory and Smoothed Particle Hydrodynamics (SPH) to predict the stability of hybrid nanofluids. Bayesian Regularization Artificial Neural Network (ANN) employed for reducing needs of extensive simulation, faster prediction of stability for new nanofluid compositions, and support optimization. By simulating Van der Waals attractions, electrostatic repulsion, Brownian diffusion, sedimentation, particle aggregation, and dispersion behavior produces a normalized stability score for ANN training under varying zeta potential, ionic strength, and nanoparticles size. A MATLAB-based SPH solver is implemented to track interparticle dynamics over time. A physics-informed dataset consisting of 5000 samples was generated using coupled DLVO and SPH simulations considering six governing physicochemical parameters: zeta potential, particle radius, Debye length, Hamaker constant, temperature, and surfactant concentration to predict hybrid nanofluid stability. Regression analysis showed a strong correlation between the predicted and actual stability scores with a coefficient of correlation $(R)$ of 0.9907. The developed model showed good predictive capability and generalization with the overall coefficient $(R^2)$ of 0.9808, root mean square error (RMSE) of 0.0234, mean absolute error (MAE) of 0.0174, and MAPE of 3.13%. The results provide knowledge to optimize the dispersion stability of Hybrid Nanofluids in Thermal and Biomedical Applications.
Zitieren
APA
Gaydhane, S. K., & Singh, V. K. (2026). ANN INTEGRATED DLVO-SPH BASED SIMULATION FOR PREDICTING STABILITY OF HYBRID NANOFLUIDS. Advances and Applications in Fluid Mechanics, 33(2), 253-278. https://doi.org/10.17654/0973468626012
Vancouver
Gaydhane SK, Singh VK. ANN INTEGRATED DLVO-SPH BASED SIMULATION FOR PREDICTING STABILITY OF HYBRID NANOFLUIDS. Advances and Applications in Fluid Mechanics. 2026;33(2):253-278. doi:10.17654/0973468626012
BibTeX
@article{Gaydhane2026,
author = {Gaydhane, Swati Kawadu and Singh, Vinod Kumar},
title = {ANN INTEGRATED DLVO-SPH BASED SIMULATION FOR PREDICTING STABILITY OF HYBRID NANOFLUIDS},
journal = {Advances and Applications in Fluid Mechanics},
year = {2026},
volume = {33},
number = {2},
pages = {253--278},
doi = {10.17654/0973468626012}
}
EUVIMED