Analysis of thermosolutal buoyancy-driven suspension comprising nano-encapsulated phase change materials using finite element method and ANN-based MLP algorithm

• Published in: Journal of the Taiwan Institute of Chemical Engineers Vol. 168; p. 105912
• Main Authors: Kouki, Marouan, Nayak, M.K., Irshad, Kashif, Mesfer, Mohammed K. Al, Danish, Mohd, Pasha, Amjad Ali, Zahir, Md Hasan, Chamkha, Ali J.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2025
• Subjects: ANN-based MLP algorithm; Double-diffusive NEPCMs; Geometrical change; Inclined H-shaped domain; Rib and Baffle; Wavy wall; Inclined H-shaped domain; Rib and Baffle; Double-diffusive NEPCMs; Wavy wall; Geometrical change; ANN-based MLP algorithm
• ISSN: 1876-1070
• DOI: 10.1016/j.jtice.2024.105912

•Thermosolutal buoyancy-driven suspension comprising NEPCMs and entropy is studied.•The inclined H-shaped enclosure is equipped with wavy walls and fins.•Forchheimer-Brinkman extended Darcy model is applied to model porous medium.•Geometrical impact on the entropy and hydrothermal features of NEPCM is analyzed.•ANN-based MLP learning algorithm is applied to predict the values of temperature and concentration. The present research problem is the investigation of thermal behavior of double-diffusive nano-encapsulated phase change materials (NEPCMs) inside a porous H-shaped wavy cavity with two baffles at the top rib. By using accessible renewable energy sources and off-peak electricity, the application of double-diffusive NEPCM in components can lower energy consumption and interior temperature variation. Such embedded H-shaped wavy cylinder with NEPCM suspension can be applied for diversified thermal uses in cooling devices and heat exchangers. The purpose of this study is to explore the effect of geometrical change on hydro-thermal and entropy behaviors of double-diffusive NEPCM suspension within a Forchheimer-Brinkman extended Darcy medium-based H-shaped cavity with wavy walls and hot baffles at upper rib. The current problem is solved by implementing finite element method (FEM) and to estimate the values of temperature and concentration, artificial neural network (ANN)-based multi-layer perceptron (MLP) is applied. The major outcomes are that the average Nusselt number whittles down by 22%, 32%, and 44% and average Sherwood number peters out by 21%, 33%, and 53% for rise in baffle length from 0.1 to 0.3, rib's size from 0.25 to 0.35 and angle from 0° to 90°, respectively. Moreover, Bejan number decreases with amplification of Raleigh number and upsurges due to augmentation of baffle length. Furthermore, mean square error for the validation data is 3.0649e-4 which shows superior ability of applied ANN-based MLP in estimating the values of temperature and concentration. Subject to proper alteration of several geometries such as baffle length, rib size, wavy wall undulation, and inclination angle, greater cooling effect of H-shaped chamber is accomplished in the presence of double diffusive NEPCM suspension applicable to lowering energy consumption and interior temperature variation thereby improving thermal performance of components. [Display omitted]