Perepi Rajarajeswari
Finite Element Numerical Simulation of Free Convection Heat Transfer in a Square Cavity Containing an Inclined Prismatic Obstacle With Machine Learning Optimization
Rajarajeswari, Perepi; Arasukumar, Thilagavathi; Anwar Bég, O.; Bég, Tasveer A; Kuharat, S.; Reddy, PBA; Ramachandra Prasad, V.
Authors
Thilagavathi Arasukumar
O. Anwar Bég
Tasveer A Bég
Ms Sireetorn Kuharat S.Kuharat2@salford.ac.uk
Lecturer
PBA Reddy
V. Ramachandra Prasad
Contributors
Prof Osman Beg O.A.Beg@salford.ac.uk
Project Leader
Abstract
ABSTRACTThe present work describes a numerical simulation of free convection heat transfer inside a square cavity containing a prismatic obstacle at various angles of inclination. The nondimensional governing equations are discretized by the finite element method and solved in the commercial software “COMSOL Multiphysics 6.1” with appropriate boundary conditions. The effect of prominent parameters on streamline, isotherm contours, and local Nusselt number profiles are depicted graphically. The control parameters are the Prandtl number and Rayleigh number (103 ≤ Ra ≤ 106). The study considers air as the circulating fluid with the Prandtl number, Pr = 0.71. The computations are conducted for the prismatic shape at four different orientations of , and . The inclination angle of the prismatic obstacle is observed to exert a significant role in the distribution of heat and momentum inside the square cavity. Furthermore, neural network approaches are used for optimizing the thermal performance of the system, via Bayesian regularization machine learning analysis and Levenberg–Marquardt algorithms. The study finds applications in solar collectors, fuel cells, and so forth.
| Journal Article Type | Article |
|---|---|
| Acceptance Date | Feb 8, 2025 |
| Online Publication Date | Feb 26, 2025 |
| Deposit Date | Feb 8, 2025 |
| Publicly Available Date | Feb 27, 2026 |
| Journal | Heat Transfer |
| Print ISSN | 2688-4534 |
| Electronic ISSN | 2688-4542 |
| Publisher | Wiley |
| Peer Reviewed | Peer Reviewed |
| DOI | https://doi.org/10.1002/htj.23315 |
| Keywords | Square enclosure; Finite Element Method; Prismatic obstacle; natural convection; COMSOL finite element software; neural network optimization; Bayesian regularization; Levenberg-Marquardt algorithm |
Files
This file is under embargo until Feb 27, 2026 due to copyright reasons.
Contact O.A.Beg@salford.ac.uk to request a copy for personal use.