Supervised Learning Algorithm to Study the Magnetohydrodynamic Flow of a Third Grade Fluid for the Analysis of Wire...

Area of Science:

• Fluid Dynamics
• Computational Science
• Materials Science

Background:

• Wire coating processes involve complex fluid dynamics, particularly magnetohydrodynamics (MHD).
• Analyzing third-grade fluids under MHD conditions presents significant computational challenges.
• Accurate modeling is crucial for optimizing wire coating quality and efficiency.

Purpose of the Study:

• To develop and validate a novel computational approach for analyzing MHD flow of third-grade fluids in wire coating.
• To implement a Levenberg-Marquardt back-propagation supervised neural network (LMB-SNN) for this analysis.
• To assess the accuracy and effectiveness of the proposed LMB-SNN solver.

Main Methods:

• Mathematical formulation of MHD flow for third-grade fluids using partial differential equations.
• Transformation of PDEs into ODEs via dimensionless parameters and suitable transformations.
• Development of a reference dataset using Adam's numerical technique for training and validation.
• Application of LMB-SNN with parallel training, testing, and validation using Mean Square Error Function (MSEF).

Main Results:

• The LMB-SNN effectively computed approximate solutions for MHD-TGFWCA across various physical parameter variations.
• Comparative analyses using MSEF, regression plots, and error histograms demonstrated the solver's high performance.
• The numerical outputs closely matched dataset values, confirming the method's precision ($$ to $$).

Conclusions:

• The Levenberg-Marquardt back-propagation supervised neural network (LMB-SNN) is a highly effective and precise solver for MHD flow of third-grade fluids in wire coating.
• The proposed computational framework offers a robust tool for analyzing complex fluid dynamics in industrial coating processes.
• The validated accuracy of LMB-SNNs paves the way for improved simulation and optimization in wire manufacturing.