Abstract

Significance: The outstanding thermal conductivity and heat transfer properties of nanofluid make them highly useful for applications in thermal engineering and other areas. With their improved effectiveness, nanofluid significantly enhance the performance of heating, cooling, and overall thermal regulation systems. Hybrid nanofluids are used in industry as heat-transport fluids in gas turbine rotators and rotating machinery. Recurrent neural network has garnered significant attention in academic research for their ability in modeling complex, nonlinear systems. Their adaptability makes them highly appropriate for advanced domains such as fluid dynamics, natural language processing, biological computing, control systems, multimedia and biotechnology, where pattern learning and recognition are critical. Purpose: This study explores heat transport in ternary hybrid nanofluid comprising capper, silver, and alumina nanoparticles between two stretching spinning disks at a constant distance. It integrates the effects of thermal radiation, joule effect, heat source, and activation energy to assess their combined influence on flow and thermal characteristics. The work further investgates the capability of a recurrent neural networks enhanced with the Levenberg-Marquardt method (RNN-LMM) to accurately model and predict these complex thermos fluid phenamena. Methodology: Through similarity transformation, the governing partial differential equations are reduced to dimensionless ordinary differntiao equations, which are then solved using the RNN-LMM. Data for the study, was obtained using the Adams numerical method and further optimized through the recurrent neural network's framework. The model was trained on eighty percent of the dataset, with ten percent allocated for testing and ten percent for validation. Performance assessment was conducted using mean squared error (MSE), regression analysis, and histogram-based error distribution, with accuracy in the range of E−3 to E−7. Graphical analysis was employed to investigate the influence of key physical parameters on velocity, temperature, and concentration fields. Finding: Results reveal that an increase in the magnetic parameter augments both velocity and temperature distributions. The Reynolds number significantly affects radial, tangential and axial velocity components, promoting overall fluid motion. Activation energy positive effect ternary hybrid nanofluid (THNF) concentration, whereas the Schmidt number and chemical reaction rate decrease it, highlighting their opposing effects. All examined factors contribute to elevated temperature profiles. The reduced MSE indicates that RNN-LMM predictions closely match true values, confirming the methods reliability and accuracy.