Designing artificial intelligence computing techniques to study heat transfer of a ternary hybrid nanofluid flow: Application of particle swarm optimization and artificial neural network
Sawan Kumar Rawat, Moh Yaseen, Manish Pant, Chandan Singh Ujarari
Abstract
This paper investigates the steady, and three-dimensional non-Newtonian and Newtonian ternary nanofluid flow toward a bidirectionally stretching surface. The non-Newtonian model is developed using the Maxwell fluid model with engine oil and the Newtonian model is developed with water as a base fluid. The analysis is done with Aluminium oxide (Al 2 O[Formula: see text] (spherical-shaped), Carbon nanotubes (CNT) (cylindrical-shaped), and Graphene (platelet-shaped) as nanoparticles. The mathematical model is articulated and it is further solved with similarity transformations. Furthermore, the analysis considers the influence of the Cattaneo–Christov model, magnetic field, and thermal radiation. A dataset is generated by adjusting relevant parameters through the bvp4c function in MATLAB. Then two different artificial intelligence computing techniques using fuzzy particle swarm optimization (FPSO) and artificial neural network (ANN) are developed to analyze the Nusselt number for Newtonian and non-Newtonian flow. The dominant values of the radiation parameter and thermal relaxation parameter have a significant impact on the HT rate, causing an increase in HT rate. The HT rate is seen to be higher for non-Newtonian TNF. The results of this model can be used in various fields including energy generation, air conditioning, nuclear reactor cooling, electronic device cooling, tissue heat conduction, desalination, crop preservation from freezing, and food processing.