Heat transfer analysis in a longitudinal porous trapezoidal fin by non-Fourier heat conduction model: An application of artificial neural network with Levenberg–Marquardt approach

Abstract

Thermal critical issues commonly occur in advanced electrical devices as a response to excessive heat generation or a loss in efficient surface area for heat exclusion. This issue can be addressed by utilizing the extended surface to improve the heat transfer performance of the devices. Thus, the present analysis is devoted to scrutinizing the non-Fourier unsteady heat transference of a trapezoidal porous fin. Levenberg–Marquardt technique of backpropagation artificial neural network (LMT-BANN) is employed here to analyze the thermal variation in the fin. The developed governing equation is the hyperbolic heat conduction equation (HHCE), which is transformed into a dimensionless partial differential equation (PDE) using dimensionless variables. LMT-BANN is employed on thermal numerical data and is developed to trace numerical approximation of fin problem using a methodology that includes testing, training, and validation. A graphical visualization of the consequences of thermal variables on the temperature field is presented. The notable evidence of this study reveals that as the magnitude of the convection factor increases, thermal dissipation through the fin gradually decreases. Further, the LMT-BANN technique has been determined to be an effective, reliable, and rapidly convergent stochastic computational solver that can be used effectively for examining the thermal model.

Keywords

• Energy transfer
• Porous fin
• Trapezoidal profile
• Non-Fourier heat flux
• Artificial neural network