Thermal analysis of flowing stream in partially heated double forward-facing step by using artificial neural network

Abstract

The regulators for thermal energy transfer, performances of heat exchangers, turbine blades subject to cooling structure, and energy storage procedures claim the use of a heated fluid with partially heated circular obstructions rooted in confined domains. Owing to such importance we consider a partially heated double forward-facing step (DFFS). To be more specific, from the inlet of DFFS, the viscous stream flows in parabolic form and the Neumann condition is implemented at the outlet. At each wall, no slip is incorporated. The mathematical formulation is constructed to narrate the flow field. The translation of the centers of mounted heated obstructions is considered in three separate situations. For every event, the strength of the Nusselt number is debated numerically. For all cases, the drag coefficient for partially heated obstruction is found a decreasing function of the Reynolds number. Besides this, for better estimation of Drag Coefficient (DC) and Lift Coefficient (LC), an artificial neural network (ANN) model with multilayer perceptron (MLP) is developed. MoD values shows that the error rates of the ANN model are very low. The findings show that the constructed ANN model can accurately predict DC and LC values with very low error rates.

Keywords

• Heat transfer
• Heated obstruction
• DFFS
• ANN model
• Hybrid meshing
• Finite element analysis