Search

Highlights: [•] Methodology using ANN and genetic algorithm to perform multi-optimization. [•] Objective-functions are annual energy consumption and summer comfort degree. [•] Optimization variables are thermophysical properties of external envelope. [•] Optimized solutions are given as a function of the climate. [•] More accurate results with multi-objective optimization than with enhanced mono-objective optimization. [ABSTRACT FROM AUTHOR]

This article deals with laminar natural convection in three-dimensional air-filled enclosures. The cavities under investigation have two opposite isothermal vertical walls and four adiabatic walls. The vertical and horizontal aspect ratios are in the range 5–80 and 0.2–5, respectively. The Rayleigh numbers vary between 103 and 105. These values match cavities typically found in vertical hollow alveolar bricks. The effects of the geometry and the Rayleigh number on the convection are discussed. Comparisons between purely two-dimensional cases and the results obtained in 3D are presented and analyzed. Nusselt number correlations as a function of the two aspect ratios and Rayleigh number are proposed, offering an efficient tool for thermal engineers and scientists to evaluate convection heat transfer in real three-dimensional enclosures, which are common systems in building construction. [ABSTRACT FROM PUBLISHER]