An artificial intelligence model for heat and mass transfer in an inclined cylindrical annulus with heat generation/absorption and chemical reaction

• Published in: International communications in heat and mass transfer Vol. 147; p. 106956
• Main Authors: Shilpa, B., Leela, V.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.10.2023
• Subjects: Artificial neural network model with Levenberg – Marquardt algorithm; Chemical reaction; Heat generation/absorption; Implicit finite difference method; Inclined annulus; Magnetohydrodynamics; Heat generation/absorption; Magnetohydrodynamics; Artificial neural network model with Levenberg – Marquardt algorithm; Chemical reaction; Implicit finite difference method; Inclined annulus
• ISSN: 0735-1933; 1879-0178
• DOI: 10.1016/j.icheatmasstransfer.2023.106956

Present research investigates MHD mixed convective two dimensional flow, heat and mass transfer in a concentric inclined annulus with zero, first and second order chemical reactions along with heat generation/absorption. The internal and external cylinders are considered as isothermal with distinct temperatures and also presumed that the upper and lower sides are adiabatic. The governing equations are solved using the quasi-linearization approach in conjunction with an implicit Crank-Nicolson type FDM to arrive at the non-similar solutions. The results indicate that as the inclination parameter augments the concentration and thermal profiles augments in the annular region. As the reaction rate increases the concentration profile enhances in the annular region, since the concentration boundary layer thickens with higher reaction rate. The temperature gradient of the fluid is significantly affected by an external heat generation parameter, which enhances the fluid temperature distribution and thermal state. The results revealed that a magnetic field can regulate the magnetic convection of conducting fluid. The 3D view of fluid flow in radial and axial direction, temperature and concentration variations for distinct angle of inclinations are also illustrated and discussed explicitly. The FDM data is used to train an ANN model that forecasts the heat and mass transfer characteristics through Levenberg – Marquardt backpropogating algorithm. For various values of significant physical parameters the heat and mass transfer rates has R values as 0.99305 and 0.99991 respectively. The derived numerical data are compared and determined to be in exceptional concordance with previous results, in the limiting sense.