Hybrid computational modeling of solar aircraft efficiency in magnetohydrodynamics Darcy–Forchheimer flow using kerosene-based SWCNT and MWCNT nanofluids
摘要
This study develops a Python-Based Supervised Artificial Intelligence (PB-SAI) framework to examine the nonlinear thermofluidic performance of a kerosene-bound single- and multi-walled carbon nanotube (SWCNT–MWCNT) hybrid nanofluid over a stretchable porous surface subjecting to solar radiation and Darcy–Forchheimer resistance. The proposed framework develops a magnetohydrodynamic (MHD) model with critical transport phenomena involving Joule heating, thermal radiation, and viscous dissipation. The governing partial differential equations (PDEs) will be reduced to a coupled set of ordinary differential equations (ODEs) via similarity transformations, using solve_bvp in SciPy to numerically solve the system. The high-fidelity numerical solutions for the velocity and temperature profiles from the governing PDEs will be used as training data for a deep feedforward neural network (DFNN) built in PyTorch, utilizing Swish activation function and Adam optimizer. The trained artificial intelligence (AI) realizes high competency with MSE up to 10–6, along with regression accuracy around 0.9966, in confidence across nonlinear regimes. There are five dimensionless parameters on the performance metrics which were examined: magnetic parameter (