A machine learning analysis for hybrid nanofluid flow between two co-axial cylinders

Area of Science:

• Fluid Dynamics
• Heat Transfer
• Nanotechnology

Background:

• Investigating heat transfer and fluid flow in confined geometries is crucial for thermal management.
• Hybrid nanofluids offer enhanced thermophysical properties over traditional fluids.
• Magnetohydrodynamics plays a significant role in controlling fluid behavior and heat transfer.

Purpose of the Study:

• To analyze steady, incompressible, heat transfer, and axisymmetric flow of a hybrid nanofluid (magnetite/multi-walled carbon nanotubes in water).
• To explore the effects of rotation, magnetic fields, and varying flow parameters on the hybrid nanofluid.
• To develop and validate a computationally efficient predictive model using artificial neural networks.

Main Methods:

• Numerical solution of the boundary-value problem using MATLAB's bvp4c solver.
• Evaluation of effective thermophysical properties via the Hamilton-Crosser mixing model.
• Development of a Levenberg-Marquardt trained artificial neural network for rapid prediction.

Main Results:

• Increased magnetic strength and Brinkman number raise temperature due to viscous and Joule heating.
• Higher Reynolds number increases the pressure gradient.
• Hybrid nanofluid exhibits 7.09% higher heat transfer and increased wall shear stress compared to single-particle nanofluids.

Conclusions:

• The hybrid nanofluid demonstrates superior heat transfer performance.
• The combined numerical and artificial intelligence approach provides accurate and efficient analysis of magnetohydrodynamic thermal systems.
• The developed neural network accurately predicts system behavior, enabling rapid parametric studies.