Developing an Artificial Neural Network Algorithm for Heat and Mass Transfer Assessment in Ternary Hybrid Nanofluid Flow

This chapter is designated to study the influence of thermal radiation and thermophoretic particle don heat and mass transport of ternary hybrid nanofluid flow in a stretching/shrinking convergent channel. The modeling expressions for the considered mathematical model are addressed with the aid of Galerkin finite element analysis and obtained results are compared with existing literature under the limiting instances. The results disclosed that an enhancement in thermal radiation term accelerates the heat transfer of the fluid. As the thermophoretic parameter enhances, the concentration decreases because the thermophoretic force draws additional particles to the surface when there is a large heat difference. An artificial neural network (ANN) is trained on simulated data using the Levenberg–Marquardt technique to identify and comprehend the underlying structures. The mass and heat transfer in the convergent channels that are stretching and shrinking are then computed using the trained neural network. This approach offers a workable replacement for traditional parametric research, enabling more precise predictions and comprehension of the behavior of complex systems. Ternary hybrid nanofluid flow has the highest mass and heat transfer rates in the stretched channel. Throughout all examined situations, the ANN algorithm's mean-squared error (MSE) values range from 0 to 0.03972. The heat and mass transfer properties of the ternary hybrid nanofluid are predicted through this algorithm. Also, the regression plot, error histograms, performance plot, and learning curves are presented for the predicted model. The research results enhance our comprehension of the properties and behavior of ternary hybrid nanofluid flow in convergent channels, which helps with mass and heat transfer design and optimization.