https://doi.org/10.1140/epjs/s11734-025-01974-5
Regular Article
Hybrid ISPH-AI simulation of oscillating tree and T-shaped fins driven double diffusion of NEPCM in an hourglass-shaped porous cavity under magnetic, radiative, and reactive conditions
1
Department of Mathematical Sciences, Princess Nourah Bint Abdulrahman University, P. O. Box 84428, 11671, Riyadh, Saudi Arabia
2
Department of Mathematics, King Khalid University, POB 960, Postal Code 61421, Abha, Saudi Arabia
Received:
21
June
2025
Accepted:
16
September
2025
Published online:
25
September
2025
This study presents a hybrid numerical–machine learning framework for analyzing and predicting double-diffusive convection of nano-encapsulated phase change material (NEPCM) suspensions inside a porous hourglass-shaped enclosure equipped with dynamically oscillating T-shaped and tree-shaped fins. This study investigates buoyancy-driven double-diffusive convection, wherein low-amplitude fin oscillations serve as a secondary perturbation without overpowering the dominant natural convection effects. A semi-implicit smoothed particle hydrodynamics (ISPH) solver is employed to capture the transient behavior of natural convection under the combined effects of magnetic fields (MHD), thermal radiation, chemical reactions, and cross-diffusion phenomena (Soret and Dufour effects). The moving fins introduce controlled sloshing perturbations characterized by amplitude and frequency, which significantly enhance thermal and solutal transport. The latent heat buffering effect of NEPCM is also incorporated via a phase-change heat capacity model. A comprehensive parametric study reveals the critical roles of Darcy number, Hartmann number, buoyancy ratio, and radiative strength in shaping the flow topology and heat/mass transfer. Additionally, a novel machine learning approach using XGBoost regression is developed to predict average Nusselt and Sherwood numbers based on time and sloshing amplitude. The trained models achieve high accuracy, with key features such as time–amplitude interaction terms identified as dominant predictors. This integration of physics-based simulation and AI modeling enables efficient surrogate prediction of complex thermal-fluid systems. The novelty of this work lies in combining multi-physics ISPH simulations with explainable AI to understand unsteady convective melting and solutal transport in porous NEPCM systems under oscillatory and reactive conditions. These findings hold strong relevance for the design of advanced thermal energy storage units, microreactors, and magneto-convective heat exchangers where dynamic flow manipulation is critical.
Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
© The Author(s), under exclusive licence to EDP Sciences, Springer-Verlag GmbH Germany, part of Springer Nature 2025
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
