Numerical simulation of boundary layer flow over a cylinder using python: a focus on homogeneous-heterogeneous reactions

• Published in: Multiscale and Multidisciplinary Modeling, Experiments and Design Vol. 8; no. 9; p. 381
• Main Authors: Taha, Talal, Khan, Muhammad Imran, Khalid, Nouman
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.09.2025; Springer Nature B.V
• Subjects: Aerospace engineering; Algorithms; Artificial intelligence; Artificial neural networks; Boundary layer flow; Boundary layers; Characterization and Evaluation of Materials; Chemical reactions; Coefficient of friction; Differential equations; Energy storage; Engineering; Environmental cleanup; Fluid dynamics; Fluid flow; Fluid mechanics; Hydrodynamics; Investigations; Laminar flow; Machine learning; Magnetic fields; Mathematical Applications in the Physical Sciences; Mechanical Engineering; Medical technology; Nanoparticles; Neural networks; Newtonian fluids; Non Newtonian fluids; Nonlinear differential equations; Numerical analysis; Numerical and Computational Physics; Original Paper; Partial differential equations; Performance evaluation; Petrochemicals; Radiation; Renewable energy; Reynolds number; Schmidt number; Simulation; Skin friction; Solid Mechanics; Solvers; Two dimensional analysis; Two dimensional flow; Viscosity; Chemical reaction; Heat and mass transfer; Artificial neural networks; Machine learning
• ISSN: 2520-8160; 2520-8179
• DOI: 10.1007/s41939-025-00973-2

Homogeneous-heterogeneous reactions are essential in many fluid dynamics and engineering applications, particularly when system behavior is governed by both bulk and surface chemical processes occurring simultaneously. To maximize performance and efficiency, it is essential to comprehend and regulate these reactions in energy storage devices, combustion systems, environmental remediation technologies, and catalytic reactors. The author analyzes two-dimensional laminar flow of a fluid with homogeneous-heterogeneous reactions along a cylindrical surface by employing Artificial Neural Networks (ANN) and Machine Learning (ML). The governing partial differential equations are converted into nonlinear ordinary differential equations through similarity transformations, and the MATLAB bvp4c solver is applied to compute a numerical solution. To assess artificial intelligence-based models, such as LMS-BPNN and machine learning approaches that use a linear regression approach, the dataset produced by the bvp4c solver is divided into training, validation, and testing sets. A combination of fitness plots, error histograms, and mean squared error (MSE) analysis is used to evaluate the model's performance and accuracy. The findings show that the concentration boundary layer grows thicker when the Schmidt number increases. Table 4 illustrates the skin friction coefficient prediction with ANN and ML. Examines the advantages and uses of Artificial Neural Networks (ANN) and Machine Learning (ML) approaches in the study of non-Newtonian fluid mechanics for both scientific and engineering purposes. The results improve practical applications in fields including petrochemicals, renewable energy, medical technology, and aerospace engineering by optimizing fluid flow and reaction dynamics.