This research investigates the transport characteristics of a ternary hybrid nanofluid (Al 2 O 3 –Graphene–CNT/water) in a rotating horizontal cylindrical annulus in the presence of the effects of MHD, thermal radiation, and irregular heat source/sink.
The sensitivity analysis indicates that nanoparticle volume fraction has the greatest sensitivity to Nusselt number (≈ 84.03%), which is the dominant factor in enhancing heat transfer, relative to other factors.
Additionally, the ANN predictions are used to construct 3D surface and contour plots of the Nusselt number, providing a comprehensive visualization of multi-parameter interactions.
The research introduces a hybrid PINN-ANN architecture for this multi-physics rotating annulus problem, providing mesh-free, high-accuracy solutions to a problem that conventional numerical methods have had convergence issues.
The analysis of ternary hybrid nanofluids flowing through a rotating annulus leads to advancements in thermal technology by providing improved cooling and energy transfer efficiency.
This research investigates the transport characteristics of a ternary hybrid nanofluid (Al 2 O 3 –Graphene–CNT/water) in a rotating horizontal cylindrical annulus in the presence of the effects of MHD, thermal radiation, and irregular heat source/sink. First, nonlinear governing PDEs are solved in a physics-informed neural network (PINN) architecture, with the L-BFGS optimizer, and afterwards an artificial neural network (ANN) is also used to predict the Nusselt number. ANN predictions are comparable to values obtained using PINN with error < 10–5, MSE = 1.56 × 10–6 and R2 ≈ 0.999999998. Velocity decreases with increasing Ha, rises with Re, and nanoparticle volume fraction, and both temperature and heat transfer rate increase strongly with the heat-source parameters, Eckert number and nanoparticle volume fraction. The sensitivity analysis indicates that nanoparticle volume fraction has the greatest sensitivity to Nusselt number (≈ 84.03%), which is the dominant factor in enhancing heat transfer, relative to other factors. Additionally, the ANN predictions are used to construct 3D surface and contour plots of the Nusselt number, providing a comprehensive visualization of multi-parameter interactions. The research introduces a hybrid PINN-ANN architecture for this multi-physics rotating annulus problem, providing mesh-free, high-accuracy solutions to a problem that conventional numerical methods have had convergence issues. The phenomena of ternary hybrid nanofluids flowing through a rotating annulus is prevalent in different fields of engineering such as turbomachinery, aerospace engineering, nuclear engineering, automotive engineering, rotary machinery, renewable energy systems, and industrial heat exchange processes. The analysis of ternary hybrid nanofluids flowing through a rotating annulus leads to advancements in thermal technology by providing improved cooling and energy transfer efficiency.
