AUTHOR=Ali Aatif , Ahammad N. Ameer , Tag-Eldin Elsayed , Gamaoun Fehmi , Daradkeh Yousef Ibrahim , Yassen Mansour F. 

TITLE=MHD williamson nanofluid flow in the rheology of thermal radiation, joule heating, and chemical reaction using the Levenberg–Marquardt neural network algorithm

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 10 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2022.965603

DOI=10.3389/fenrg.2022.965603

ISSN=2296-598X

ABSTRACT=Various studies have been conducted on the topic of predicting the thermal conductivity
of nanofluids. Here, the thermal conductivity of nanofluid is determined using Artificial Neural
Networks (ANNs) since this approach is both rapid and accurate, as well as cost-effective. To
forecast the thermal conductivity of MHD Williamson nanofluid flow through a vertical sheet,
feed-forward neural network with various numbers of neurons has been evaluated and the best
network based on the performance is selected. The fluid model incorporates the effects of Joule
heating, heat generation absorption, thermal radiation, and a chemical reaction (MHD-WNFHGA). A combination of heat radiation and reactive species improves the energy and solute
profiles. The magnetic Reynolds number is assumed so small therefore the generated magnetic
field has no effect. A postulate of similarity variables are used to convert the physical model in the
form of nonlinear partial differential equations to ordinary differential equations system. A
supervised Levenberg–Marquardt backpropagation algorithm possesses multilayer perceptron is
used for training the network, which is one of the top algorithms in machine learning. The bvp4cnumerical technique is adopted to build the datasets for the construction of a continuous neural
network mapping. Flow, energy, and concentration profiles of the fluidic flow are constructed by
adjusting several physical quantities such as Williamson parameter, thermal radiation parameter,
magnetic parameter, Eckert number, Darcy number, Brownian motion and thermophoresis
parameter. Analytical techniques such as error histogram graphs and regression-based statistical
graphs are used to examine the accuracy of a suggested method. It's been found that the LevenbergMarquardt Backpropogation neural network mappings' derivation, convergence, authentication, and consistency have been proven.