Vol.26, No. 2, 2026, pp. 266–275
https://doi.org/10.69644/ivk-2026-02-0
266

ANALYSIS OF THE MACHINE LEARNING APPROACH TO INVESTIGATE THE HALL AND SORET-DUFOUR EFFECTS OF WILLIAMSON FLUID OVER A STRETCHING CYLINDER

Vineet Kumar Chaurasiya1* , Manish Tanwar2 , Ruchika Mehta2

1) Department of Mathematics, Institute of Applied Sciences and Humanities, GLA University, Mathura, Mathura-281 406, Uttar Pradesh, INDIA

V.K. Chaurasiya https://orcid.org/0000-0002-8883-0687 , *email: 1990.vineet.1990@gmail.com

2) Department of Mathematics and Statistics Manipal University Jaipur, Jaipur, Rajasthan, INDIA

M. Tanwar https://orcid.org/0009-0006-2819-8890 ; R. Mehta https://orcid.org/0000-0002-2156-4943

 

Abstract

This work explores how non-Newtonian Williamson fluids behave when flowing around stretching cylinders under magnetic field influence. We examine several important physical factors - Hall currents, thermal production, varying thermal conductivity, and convective heating effects - all within a radially applied magnetic field. The study also accounts for thermophoretic forces, Brownian particle movement, and Soret-Dufour effects that connect heat and mass transfer processes. These combined factors make the study particularly relevant for industrial applications using nanofluids. To solve this complex problem, we transform the original nonlinear partial differential equations using similarity methods and computed solutions numerically with MATLAB's bvp4c solver. Our results provide detailed insights into how different parameters affect fluid velocity, temperature, and concentration distributions. We find that Soret and Dufour effects play a major role in heat-mass interactions. The research integrates machine learning-enhanced Multiple Linear Regression (MLR) models developed to predict fundamental transport phenomena in fluid dynamics applications. Three distinct models are established for skin friction coefficient CfRe0.5, NuxRe–0.5, and ShxRe(–0.5} using comprehensive statistical analysis. All models demonstrate exceptional predictive capabilities with R2 > 0.90 and MAPE values below 12 %, confirming their statistical significance and suitability for industrial fluid dynamics applications.

Keywords: • Williamson fluid • Hall effect • radiation effect • Soret-Dufour • BVP4C • multiple linear regression (MLR)

full article (1.42 MB)