Intelligent neural computing to investigate the heat and mass transmission in nanofluidic system between two rotating permeable disks: Supervised learning mechanism

Abstract

The prime objective of the present study is to investigate the effectiveness and accuracy of a single-trained artificial neural network. The Levenberg-Marquardt Backpropagated networks are tested for heat and mass transmission in magnetized hybrid nanofluid flow between the rotating permeable system. The reference data to generate different cases for distinct scenarios has been obtained using the Adams method in Mathematica using the ND-Solver function. Additionally, the system of highly nonlinear PDE's is achieved and transformed into ODE's. The effect of body forces such as thermophoresis particle diffusion and Brownian motion are incorporated in the hybrid nanofluid system. Configuration under observation is permeable and under constant impact of the magnetic field. Rosseland thermal radiation approximation relation is utilized to investigate the linear impact of radiation on velocity and temperature profile. The accuracy and effectiveness of the proposed ANNLMB are depicted with performance demonstrations. Mean square error, histogram error, and regression plots are generated for all the scenarios to discuss the varying impact of different key study parameters on the performance of the proposed ANNLMB. Furthermore, the generated reference data is distributed in the following manner 82 %, 9 %, and 9 % for training, testing, and validation, respectively.