Simulation of rotating Boger hybrid nanofluid flow with nanoparticles between concentric cylinders using Morlet-Wavelet neural network analysis

In this study, the Levenberg-Marquardt algorithm combined with a backpropagated artificial neural network (LMS-BANN) is employed to investigate the steady, incompressible flow of a Boger nanofluid between two closely spaced symmetrical cylinders (BFCC). The research compares the effects of single and hybrid nanoparticles on velocity, pressure, and thermal distribution. To implement LMS-BANN, the system of partial differential equations (PDEs) governing fluid dynamics is converted into a system of ordinary differential equations (ODEs) using suitable transformations. The reference dataset for LMS-BANN is generated by numerically solving these ODEs with the BVP4C method i.e Boundary Value Problem, 4th-order, collocation method. The study examines how variations in physical parameters influence the velocity and temperature profiles, utilizing regression analysis, training processes, and mean square error (MSE) graphs to evaluate and validate LMS-BANN's performance. The accuracy of the BFCC solution approximation with LMS-BANN is assessed through validation, training, and testing phases. The LMS-BANN model reported a mean square error (MSE) as small as 1.3134E−10 and practically very accurate in the prediction of the flow of fluid. Also, regression values peaked at R = 1, which displays the outstanding work of the model.

Перевод пока недоступен