ANN-based thermal analysis of 3D MHD hybrid nanofluid flow over a shrinking sheet via LMA

The study accounts for the heat transfer by a 3D mixed convection magnetohydrodynamic (MHD) hybrid nanofluid (HNF) flow along a shrinking sheet using an engine oil-based fluid emulsified with [Formula: see text] nanoparticles. The governing nonlinear equations are simplified through similarity transformations and solved numerically using MATLAB's bvp4c method. An artificial neural network (ANN), trained using the Levenberg-Marquardt algorithm (LMA), is employed to learn from numerical data with high accuracy and speed, thereby predicting both flow and temperature fields. Compared to conventional numerical methods, the ANN-based approach offers improved computational efficiency. The effects of dominant parameters, Grashof number, thermal radiation, magnetic field strength, nanoparticle volume fraction, and internal heat generation, on velocity and temperature profiles are analyzed. Heat transfer improves with increased nanoparticle volume fraction and Grashof number, while magnetic fields and slip reduce fluid velocity. ANN predictions exhibit close agreement with numerical results, confirming the model's accuracy. Specifically, the LMA-ANN model achieved a regression coefficient of [Formula: see text] and mean squared errors as low as [Formula: see text] for velocity and [Formula: see text] for temperature. These results demonstrate the model's effectiveness in capturing nonlinear thermal-fluid behavior with minimal error. The novelty lies in integrating ANN with traditional numerical techniques to simulate hybrid nanofluid flows, offering valuable insights for applications in electronic cooling, energy systems, and thermal process optimization.