Efficient Thermal Performance Prediction of Triple-Diffusive Tangent Hyperbolic Nanofluid Flow with Activation Energy and Chemical Reaction Using Computationally Aided Advanced Neural Networks

Abstract

Neural networks are frequently employed to address technical and real-world challenges due to their effectiveness and versatility. The current study investigates the intricate interaction between mixed convection, magnetohydrodynamic effects, and tangent hyperbolic nanofluid flow across an extended medium. It uses a sophisticated Gaussian neural network integrated with a hybrid Cuckoo Search method. The study examines and documents the behavior of model physical factors, revealing a direct correlation between buoyancy forces and flow dynamics, as well as an inverse correlation between solute transportation mechanisms and fluid temperature. Notable relationships were discovered with the Prandtl, Schmidt numbers, Soret, Brownian motion, and thermophoresis parameters. Also, water is treated as the base fluid flowing over a stretchable surface, incorporating TiO₂ nanoparticles. The validity and precision of the suggested procedure are demonstrated by the essential statistical indicators that are emphasized, such as the Absolute Errors range as ${10}^{-12}$to ${10}^{-05}$, the Mean Squared Errors range as ${10}^{-07}$to ${10}^{-02}$, the Theil's Inequality Coefficient range as ${10}^{-07}$to ${10}^{-01}$, and the Error in Nash-Sutcliffe Efficiency values range as ${10}^{-15}$to ${10}^{-05}$. By implementing the recommended strategy, numerical computations become considerably more precise and accurate. In several fields, including energy production, environmental engineering, and chemical processing, ensuring optimal development performance is crucial.