I need to determine local heat transfer field in CFD (PHOENICS). After introducing it as a new variable obtained from the wall function, I am not really sure from where should I read it:
- thermal boundary layer, as generally heat transfer is in the function of bulk and surface temperature,
- side of the first cell above the object,
- or the surface of the object?
One thing that worked for me was to extrapolate the temperature distribution up to the wall. Now you will have both the local wall temperature and bulk temperature to calculate local heat transfer coefficient.
A good esteem can be given by the local Nusselt Number, Defined as the ratio between the temperature gradient at the wall, and the difference between the wall temperature and the mean bulk temperature. This quantity has the dimension of m^-1, you can non-dimensionalize it by multiplying it by the pipe diameter or another significant length.
Look at publications of me and other coauthors.
https://www.researchgate.net/publication/269146979_AUTOMOTIVE_COMPACT_SURPERCHARGE-AIRINTERCOOLER_NUMERICAL_ANALYSIS?ev=prf_pub
https://www.researchgate.net/publication/223379736_Numerical_analysis_of_a_cross-flow_compact_heat_exchanger_for_vehicle_applications?ev=prf_pub
You will find that you have to apply the definition of convection heat transfer coefficient.
Conference Paper AUTOMOTIVE COMPACT SURPERCHARGE-AIRINTERCOOLER NUMERICAL ANALYSIS
Article Numerical analysis of a cross-flow compact heat exchanger fo...
I agree with Andrea above. Go to the definition of local heat transfer coefficient - from the balance between diffusive heat flux at the wall (thermal conductivity multiplied by the temperature gradient at the wall) and convective heat flux defined by the Newton cooling law as the temperature difference between wall and mean bulk temperature. From this equality you define heat transfer coefficient. So the only problem is how to calculate temperature gradient - this is an issue of numerical mathematics.
I would like to express my gratitude to you all on your answers. The thing which was confusing me was the difference between heat transfer coefficient and wall heat transfer coefficient. After some reading, calculations and your help now I am able to extract the results.
I would advise in using the adiabatic wall temperature of the wall point as the reference free stream temperature for the heat transfer coefficient. If you have adiabatic wall condition imposed , you already have this number. If not you could obtain it from a separate run. Then, using local heat flux and local wall temperature you could calculate the heat transfer coefficient. The determination of the reference free stream temperature is of extreme importance to find a theoretically correct heat transfer coefficient. WE use this approach in our experimental research turbine with great effectivenes. I am attaching a paper describing the specific effort.
Article Determination of Casing Convective Heat Transfer Coefficient...
In my current research (heat transfer coefficient for the perforated plate), I have obtained the temperature of the PP, the free stream velocity, temperature and I have calculated heat flux and Nusselt criteria equation. Now I would like to calculate local heat transfer with means of CFD. The boundary conditions are obtained average temperature (the average readings of thermocouples) for the PP, free stream velocity and its temperature. For the validation are used heat flux obtained by measurements and air temperature on the outlet. At the moment, the difference between measurements and CFD is +- 15%.
Mr. Mladen Aleksa Tomić
will you please like to share that how you calculated the heat transfer coefficient in CFD??? In a lit bit detail please as it is my first exposure to Fluent.
Best Regards
Similar to the experimental research. In the first phase of my CFD research, I needed to calculate the heat transfer coefficient for a single perforated plate. I have set a constant temperature for the solid block - a 3D model of a p. plate. Then, I have set the inlet temperature of the air and was monitoring the outlet air temperature. The difference of these temperatures gave me the exchanged heat (mass flow*heat capacity*(Toutlet-Tinlet)). Further, according to the Newton's law, the transferred heat is equal to the heat transfer coefficient*difference between temperatures*heat transfer surface. In my early papers, I was using the difference between the plate temperature and the inlet air temperature, what was the kind of error. In my later papers, I have been using the difference between the average air temperature (Tinlet+Toutlet)/2 and the average temperature of the solid (per. plate in my case). Once, You extract the heat transfer coefficient, and You could thing of finding the Nusselt criteria.
Of course, You will always need the experimental data for the comparison, and/or others authors research resalts.
I hope I have helped in some way. You could also check some of my papers.
Regards,
Mladen
Hi Mladen
the local heat transfer coefficient on a surface is defined from the local temperature gradient on the surface (using the Fourier law for the conduction heat transfer on the liquid side with the coefficient of thermal conductivity of liquid), devided by the temperature difference (surface-liquid). So from the temp profile you have to calculate temperature gradient first. I don´t know hpw Phoenics calculate the heat transfer coefficient, but must do it a similar way I describe above.
Miroslav
No, I was not extracting the local heat transfer coefficient. I was looking for the global, average heat transfer coefficient. In order to make me things easier, I have assumed the constant wall temperature of the solid.
In the experiment I was using the average temperature of the sections of the plate.
Yes I understand, but from the local HTC you can obtain the average one by integrating the local one along the surface and making the average.
Yes, I have also understood that, but I seemed to me this was easier because I could set up the constant temperature. I have also done in ANSYS local heat transfer research and compared it to the obtained one for the separate heat transfer surfaces of the p. plate. Also, the local heat transfer coefficent was proportional to the experimentaly obtained ones (~velocity^obtained exponent).
- Badges
- Science topic