A novel machine learning approach for numerical simulation on the hybrid nanofluid flow past a converging/diverging channel: Properties of tantalum and alumina nanoparticles

Convergent and divergent channels have practical uses in the manufacture of fibres and glass, for the production of plastic sheets, the control of molten metal flows, and casting of metals. Also, hybrid nanofluids are being explored as possible working fluids for solar collectors and other high heat flux systems because of their outstanding heat transfer abilities. The present study focuses on exploring the effect of quadratic thermal radiation on the hybrid nanoliquid stream via a convergent/divergent channel, considering the impact of homogeneous-heterogeneous chemical reactions with the suspension of tantalum and alumina nanoparticles on the liquid stream. The governing partial differential equations of the present problem are transformed into dimensionless ordinary differential equations with the help of similarity variables. Further, the resultant equations are solved using the finite element approach. Also, the Laguerre Polynomial-based Physics Informed Neural Network (L-PINN) model is adopted to analyze the fluid flow, heat, and mass transfer characteristics. The significance of pertinent physical parameters on the field variables is depicted with graphical representations. Turbulence, caused by chaotic fluid motion, increases energy dissipation, which can limit the channel's effective flow velocity in stretching convergent channels. A rise in the Reynolds number in a divergent stretchable and shrinkable channel raises the temperature due to improved convective heat transfer.