Heat Transfer and Flow Analysis of Natural Convenction aroud a Toroidal Surface Incorporating Radiation Effects a Numberical Approach
Loading...
Date
item.page.authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
vi
newlineABSTRACT
newlineEfficient thermal management remains a cornerstone of modern engineering systems, particularly in the applications of toroidal shaped objects where fluid flow and heat transfer behavior are of utmost important. The objective of this thesis is to explore comprehensive numerical investigation of conjugate heat transfer, entropy generation and effect of rotation on heat loss in a solid torus. Also, the computational fluid dynamics (CFD) methods have been combined with contemporary machine learning (ML) approaches to advance both physical understanding and predictive capabilities. The simulations were performed in ANSYS Fluent using the Finite Volume Method, focusing on an aluminum torus maintained at a constant surface temperature and exposed to ambient air. The analysis accounted for the simultaneous influence of natural convection and thermal radiation, with key parameters including aspect ratio, Rayleigh number, and surface emissivity which were systematically varied to evaluate their impact on thermal performance. The study also employed the Taguchi optimization method to assess the relative influence of input parameters on thermal output. In addition to analyzing heat transfer characteristics, the research explored thermodynamic irreversibilities by quantifying entropy generation arising from both heat transfer and fluid friction. The parametric study included variations in surface temperature, aspect ratio, and Rayleigh number to determine optimal configurations that minimize entropy generation and enhance thermal efficiency. Furthermore, to complement the physics-based analysis, a machine learning framework have been also developed using data generated from computational simulations. Six different regression algorithms were trained to predict the average Nusselt number, serving as a surrogate model for rapid estimation of heat transfer performance. This data-driven approach demonstrated the feasibility of integrating machine learning models into thermal system design, enabling faster evaluations without compromising reliability. The study is further extended to mixed convection scenarios by introducing rotational effects in the torus. The impact of rotation on fluid flow and thermal exchange enhancement was examined. Overall, this thesis presents the integrated methodology combining detailed CFD modeling, thermodynamic analysis, and machine learning prediction which provides robust insights and practical strategies for optimizing heat transfer and reducing entropy generation in toroidal shaped objects prone to engineering applications.
newlineKeywords: Torus; CFD; Conjugate heat transfer; Natural Convection; Mixed convection; surface radiation; DO model; Emissivity; Rayleigh number; Nusselt number; Bejan number;
newlinevii
newlineOptimization; Taguchi; Entropy generation; Heat transfer irreversibility, Fluid friction irreversibility; Machine learning; XGBoost; Rotating solid torus.
newline