Thermal plume intensification and double-diffusive transport in sawtooth-finned non-Newtonian enclosures with dual-ANN prediction
Year: 2026
Type: article
Abstract: Advanced thermal engineering increasingly demands compact geometries that intensify heat and mass transport for energy storage, electronic cooling, and high-performance heat exchangers. This study introduces a centrally heated and concentrated triangular sawtooth fin within a non-Newtonian enclosure, offering a geometry-driven strategy to enhance thermo-solutal convection and improve next-generation thermal management efficiency. Three fin orientations are considered: Backward-Inclined Asymmetric Sawtooth Fin (BIASF), Symmetric Triangular Sawtooth Fin (STSF), and Forward-Inclined Asymmetric Sawtooth Fin (FIASF). The governing equations are solved using the finite element method with the Newton–PARDISO solver. Artificial neural network models based on Levenberg–Marquardt and Bayesian Regularization algorithms are also developed to predict the average Nusselt and Sherwood numbers. The results show that increasing the Rayleigh number intensifies buoyancy-driven circulation, which enhances both heat and mass transfer inside the enclosure. Among all configurations, the STSF case provides the strongest thermo-solutal transport performance due to its symmetric triangular structure, which promotes stronger flow interaction near the sawtooth edges. Compared with BIASF and FIASF, the STSF configuration improves the average Nusselt number by approximately 3.75–4.61% and the average Sherwood number by about 3.57–4.34%. BIASF and FIASF exhibit almost identical heat and mass transfer behavior. The local Nusselt and Sherwood number distributions show strong oscillations near the sawtooth tips, confirming that geometric irregularities play an important role in enhancing thermal and concentration transport. The ANN models show high prediction accuracy with low mean square error and correlation coefficients close to 99%.
Language: English
Subfield: Computational Mechanics
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,560
OpenAlex ID: W7222014344