Numerical Study Of Turbulence Nanofluid Flow To Distinguish Multiphase Flow Models For In-House Programming
MetadataShow full item record
Fluid flow with particles are found in many engineering applications such as flows inside lab-on-a-chips and heat exchangers. In heat exchangers, nanofluids or base fluids mixed with nanoparticles are applied to be used as the working fluid instead of the traditional base fluids which have low thermal physical properties. The nanoparticle diameters are in the range from 1 to 100 nanometers are mixed with the traditional base fluids before they are applied inside the heat exchangers and the nanofluids have been proved continually that they enhance heat transfer rates of the heat exchangers. Turbulent and laminar nanofluid flows have shown different enhancements in different conditions. This work focused on comparing different turbulent nanofluid simulations which used the computational fluid dynamics, CFD, with different multiphase models. The Realizable k-s turbulence model coupled with three multiphase models; Volume of Fluid (VOF) model, Mixture model and Eulerian model, were considered and compared. The heat exchanger geometry in the work was rectangular as in the electrical device application and the nanofluid was a mixture between A1203 and water. All simulated results, then, were compared with experimental results. The comparisons showed that numerical results did not deviate from each other but their delivered-time consumptions and complications were different. If one develops his own code, Eulerian model was the most complicated while Mixture model and Eulerian model consumed longer performing times. Although the Eulerian model delivered-time consumption was long but it provided the best results, so the Eulerian model should be chosen when time consumption and errors play important roles. From this ordinary study, the first significant step of in-house program developments has started. The time consumption still indicated that the high performance computers should be selected, and properties obtained from the experimental studies should be imported to the simulation to increase the result accuracy.