Physics-informed deep learning study for MHD particle-fluid suspension flow with heat transfer in porous annular-sector duct

Abstract

Thermal enhancement remains a critical requirement in different engineering applications. Many factors can affect the efficiency of the techniques used for this aim. The purpose of this study is to investigate the impact of particle-fluid suspensions with heat transfer through porous annular-sector duct on enhancement techniques and address the potential application of deep learning to suspension problems. The analysis is focused on the fully developed region of the forced convection flow. Thermal and rheological properties of particle-fluid suspensions were studied using physics-informed neural networks exploiting transfer learning capabilities for making parameter analysis. Another finite element solution was introduced as a measure of accuracy and to support our findings. Results were prepared in a comparative manner for both solvers including contour plots, tabular, and two dimensional figures. The average Nusselt number and friction factors were calculated for different cases to investigate the value of the thermal performance factor. Our results indicate the downside of suspensions on thermal enhancement and their negative impact on other techniques.