SANN modeling of a second-grade hybrid nanofluid flow over a Darcy–Forchheimer rotating disk

Effective thermal management of rotational motion and transport in porous media plays a major role in the design of efficient cooling systems, energy conversion and mechanical devices. This present work examines the thermo-mass transfer behaviors of radiative Rivlin-Ericksen hybrid nanofluid over a rotating disk with porous medium characterized by the Darcy-Forchheimer. In this proposed system, carboxymethyl cellulose-water (CMC) is used as base fluid in combination with C u and A l 2 O 3 nanoparticles for the augmentation of heat transfer effectiveness. In present consideration, we are going to develop a mathematical model assuming steady, axisymmetric flow, porous medium resistances, inertial effects, velocity slip, thermal radiation, heat generation/absorption, thermal stratification, chemical reaction, and solutal stratification effects. Finally, to obtain the results, the present mathematical equations, based on the non-linearity to ordinary differential equations in conjunction with Von Kármán similarity transformations, were solved numerically employing the bvp4c solver MATLAB. A Statistical Artificial Neural Network (SANN) model is prepared based on the data generated from obtained numerical results. Numerical findings are presented with the help of comparison of our predicted results from SANN model based on different criteria like: Regression, MSE, RMSE and Correlation coefficient. Comparative result validates that SANN models are reliable. Results signify that thermal field increase due to increase in radiation and heat generation. Also, the reduction in the velocity profiles occurs due to large value of porous resistance and inertial effects. Finally, the proposed C u − A l 2 O 3 /CMC-water hybrid nanofluid demonstrates more effectiveness in the thermal analysis. Our research signifies that developed hybrid nanofluid configuration combined with certified SANN models could offer strong potential for accurate thermal management as well as heat transfer prediction in various fields of science and technology, viz, rotary-machinery and energy appliances. In the current set-up, it is suggested that the C u − A l 2 O 3 /CMC-water produces higher Nusselt and Sherwood numbers than the C u − A l 2 O 3 /CMC-water, for the range of parameters that were sampled. Within the training domain, the SANN surrogate replicates the bvp4c reference solution with a high degree of interpolation accuracy.