Artificial neural network and CBS-FEM techniques for mixed convection in lid-driven tank heated by triangular fins and filled with permeable medium: Two-energy equations model

Abstract

Background

Two equations are proposed to simulate the thermal fields, and two formulations are introduced to examine the irreversibility of the mixed convection within lid-driven tanks. The forced flow is due to the movements of the upper irregular edge while the buoyancy-driven flow is due to bottom heated triangular fins. A permeable medium fills the domain while heat generation is considered for the fluid- and solid phases. Besides, the work suspension is water-based hybrid nanofluids and the other components are Cu and Al₂O₃. Furthermore, an inclined Lorentz force takes place in the flow area.

Methods

The Characteristics-Based Split (CBS) algorithm with semi-implicit technique is applied for the gradients of the pressure and the Galerkin Finite Element Method (FEM) is used to solve all the governing equations. Also, an effective ANN analyses are performed to predict some important physical quantities such as horizontal and vertical velocity components.

Significant findings

The case where the upper irregular edge moves from left to right (Case 1) gives a higher horizontal velocity compared to Case 2 (the upper movement has the opposite direction of Case 1). The higher values of the interface heat transfer coefficient cause the buoyancy-driven flow to be dominant compared to the forced flow.