Numerical and artificial neural network-based study of magneto-thermal flow in ternary nanofluids over a convectively heated needle
Muhammad Idrees Afridi, Hina Gul, Saima Riasat, Umair Khan, Muhammad Sadiq
Abstract
Fluid flow models are fundamentally dependent on artificial intelligence (AI) to enhance the precision of simulations, facilitate more efficient model optimization, and achieve dependable fluid flow systems. The optimization of fluid flow efficiencies can be realized through the application of artificial neural networks (ANNs), which represent a contemporary technological advancement. An optimization technique identified as Levenberg-Marquardt backpropagation artificial neural network, commonly referred to as LMBPANN, proficiently addresses the complexities associated with nonlinear fluid flow. Owing to the significant implications of ternary nanofluid flow, a thin movable needle passes through various fields. The present attempt has been made to compare the ternary hybrid nanofluids and the regular nanofluids. The study aims to explore the variations in velocity and temperature profiles across various pertinent parameters. Three distinct types of nanoparticles, such as aluminum oxide, titanium dioxide, and magnetite nanoparticles, have a mixture of 20 % water and 80 % ethylene glycol. Using ternary hybrid nanofluids (THNF's) and nanofluids, a comparative evaluation was performed on the fluid flow passing over a two-dimensional thin moving needle. The bvp4c technique is utilized to obtain the numerical results of the model. It is detected that the velocity of fluid reduces with a growing magnetic field. Additionally, it is examined that as the thermal relaxation parameter increased, the heat transfer rate decreased. The study concludes that ternary hybrid nanofluids perform better overall than nanofluids.