|
[Sponsors] |
May 20, 2016, 14:37 |
Nu Empirical Correlations
|
#1 |
New Member
Join Date: May 2016
Posts: 9
Rep Power: 9 |
Hi.
I'm new user of Ansys Fluent code and I want to simulate a turbulent convective flow in rectangular duct (Re=7000). I need your help to : 1. How to calculate Nu number in Fluent? 2. What should I use to compare my results : Dittus-Boelter equation or Gnielinski correlation. Thank you. |
|
May 21, 2016, 12:17 |
|
#2 |
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,675
Rep Power: 66 |
Do the pipe flow tutorial.
There is already a variable programmed for Nusselt number. You just need to specify all the reference properties in the reference pane. Note however, that those empirical correlations are based on a bulk temperature. To get the bulk temperature in Fluent/Ansys, will cause pain in some parts of your body. You must specify the reference temperature explicitly in Fluent (again, this is in the reference pane), and it is generally a constant. |
|
May 21, 2016, 13:14 |
|
#3 |
New Member
Join Date: May 2016
Posts: 9
Rep Power: 9 |
Thank you "LuckyTran" for your reply.
The details of my simulation are shown in the attached picture. Re = 7000 Tair = 298 K, at this temperature the property of air: Density = 1,1845 kg/m3 Dynamic Viscosity= 1,8444x10-5 wich gives a velocity Uinlet = 0,109 m/s 1. You mean the fluent tutorial or that of Cornell university? 2. What is the temperature value that I have to give at the channel outlet? 3. How to calculate bulk temperature in Fluent? Thank you in advance. |
|
May 21, 2016, 13:26 |
|
#4 |
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,675
Rep Power: 66 |
1. Do any pipe flow tutorial. Either the one from Fluent or any pipe flow tutorial you can find. This is pretty much the most basic of all well-defined fluid flow problems. Just pick a tutorial that has instructions, pictures, or videos on what you need to do so that you can quickly learn how to do it.
2. It's not an outlet temperature, it's a backflow temperature at the outlet. Yes, it's very important in Fluent to actually read the words. This temperature is only used if there is reversed flow, which there shouldn't be. So put in any value that is sensible. 3. The short answer is you can't. The reason is, because the definition of bulk temperature is hard to generalized onto a generalized grid. Analytically it sounds simple, just integrate everything. But try discretizing it and then it quickly becomes an impossible problem. |
|
May 21, 2016, 13:35 |
|
#5 |
New Member
Join Date: May 2016
Posts: 9
Rep Power: 9 |
I used this tutorial to calculate bulk temperature:
https://www.researchgate.net/file.Po...01441967623533 I give a value of the outlet temperature between 298 and 340 or just 298? |
|
May 21, 2016, 13:42 |
|
#6 |
New Member
Join Date: May 2016
Posts: 9
Rep Power: 9 |
I Select inlet under "Compute From" in reference value panel and the temperature = 298 K.
I keep this value or I take a temperature between 298 and 340? |
|
May 21, 2016, 13:49 |
|
#7 |
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,675
Rep Power: 66 |
I realize that is a Fluent document. But the person who wrote that tutorial clearly has a very limited understanding on what bulk velocity and temperature are. Bulk quantities are really only meaningful in internal flows, not external flows. Fluent has its roots in chemical engineering anyway, so I'm not too surprised that their employees would be confused on thermo-fluids-heat transfer.
But the tutorial does show how much of a pain it is to use local reference values as opposed to constant values. The tutorial uses a bunch of point monitors (none of these will give you a bulk averaged temperature). You need to be performing mass-weighted averages of static temperature to get the bulk temperature. This requires creating cross-sectional slices (at every grid point). But in a generalized unstructured grid, the slices will intersect regions will no nodes and you will end up needing to interpolate the adjacent cell values onto the virtual nodes of the slice. So your bulk temperature calculation which should have been a direct calculation now becomes an approximation (an interpolated value) of an already make-believe quantity, which makes the utility of the bulk temperature a lot less meaningful. There are other ways of course for approximating the bulk temperature, using curves-fits, etc. But again, there is no straightforward way to do it. |
|
May 21, 2016, 13:58 |
|
#8 |
New Member
Join Date: May 2016
Posts: 9
Rep Power: 9 |
thank you for the time you devote to meet me, that's very nice of you.
but should I leave the temperature value = 298 in reference value panel? |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
two point correlations on unstructured grid | LuckyTran | Visualization & Post-Processing | 0 | January 14, 2015 10:30 |
Trans SST Correlations by UDF | sifou | FLUENT | 0 | December 4, 2011 17:29 |
Turbulent flow in smooth rectangular duct - Numerical vs. Literature correlations | marluc | FLUENT | 0 | July 7, 2011 09:23 |
UDF convective heat transfer correlations | sarah | FLUENT | 0 | March 29, 2007 09:52 |
Reynolds stresses without velocity correlations? | Phil | Main CFD Forum | 1 | October 2, 2006 04:29 |