Nusselt and friction factor validation in turbulent flow

me.po · April 27, 2015, 16:52

Hi
I've tried to simulate a classical problem of flow and heat transfer in a circular tube with constant heat flux, radius of 10mm and 60mm length. I've used the periodic boundary condition, K-epsilon model with st. wall function and the Reynolds number between 10000 to 40000. When I use the finer mesh, the friction factor and Nusselt increase drastically and get distance from the analytical results

. This problem is more serious for low Reynolds (10000<Re<20000). I've checked several grids from 22000 to 1 billion cells, both fine and coarse, with boundary layer mesh (first layer at 10-6 m and I've also used adaption. I've even solved the problem with non-periodic condition and fully developed profile but the problem still remains. I checked the mesh with laminar velocities (and laminar model) and the results were consistent with analytical results (less than 3% error).
Please guide me to solve this problem.

LuckyTran · April 27, 2015, 22:08

This is exactly why validation & mesh sensitivity studies are always necessary.

Most likely a mesh quality problem. If your grid was automatically generated, it's probably related to increased grid skewness when you went to the finer grid.

me.po · April 28, 2015, 03:23

Thanks,
What's your suggestion???

buidu · April 28, 2015, 04:11

The turbulence intensity may have influence on the convective coefficient. The empirical equations are on the assumption that TI is very low. You can find it on textbook. From you simulation, you can also notice it.
Perhaps you can try to run your case on lower TI.

me.po · April 28, 2015, 05:48

Quote:

Originally Posted by buidu

The turbulence intensity may have influence on the convective coefficient. The empirical equations are on the assumption that TI is very low. You can find it on textbook. From you simulation, you can also notice it.
Perhaps you can try to run your case on lower TI.

Thank you for your advice,
Yes, it's too high. My turbulence intensity is Min 50% (at core region) and Max 91% (at wall). How can I control TI in periodic domain?

buidu · April 28, 2015, 13:11

Quote:

Originally Posted by me.po

Thank you for your advice,
Yes, it's too high. My turbulence intensity is Min 50% (at core region) and Max 91% (at wall). How can I control TI in periodic domain?

You can control your TI in the setting of inlet.

Regards

me.po · April 28, 2015, 13:42

But that wasn't the cause of the error. when you use periodic conditions, there is no longer any inlet or outlet and flow is fully developed. I checked the problem with finer mesh with very good quality and the problem still remains!!!!!!!!!!!!!!!!!!

vaibhavhiwale · July 24, 2023, 04:16

Did this problem is solved? im also facing the same problem

April 27, 2015, 16:52	Nusselt and friction factor validation in turbulent flow	#1
me.po New Member Mehdi Join Date: Nov 2014 Posts: 4 Rep Power: 12	Hi I've tried to simulate a classical problem of flow and heat transfer in a circular tube with constant heat flux, radius of 10mm and 60mm length. I've used the periodic boundary condition, K-epsilon model with st. wall function and the Reynolds number between 10000 to 40000. When I use the finer mesh, the friction factor and Nusselt increase drastically and get distance from the analytical results . This problem is more serious for low Reynolds (10000<Re<20000). I've checked several grids from 22000 to 1 billion cells, both fine and coarse, with boundary layer mesh (first layer at 10-6 m and I've also used adaption. I've even solved the problem with non-periodic condition and fully developed profile but the problem still remains. I checked the mesh with laminar velocities (and laminar model) and the results were consistent with analytical results (less than 3% error). Please guide me to solve this problem.

April 27, 2015, 22:08		#2
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,763 Rep Power: 66	This is exactly why validation & mesh sensitivity studies are always necessary. Most likely a mesh quality problem. If your grid was automatically generated, it's probably related to increased grid skewness when you went to the finer grid. me.po likes this.

April 28, 2015, 04:11		#4
buidu New Member buidu Join Date: Dec 2014 Posts: 27 Rep Power: 13	The turbulence intensity may have influence on the convective coefficient. The empirical equations are on the assumption that TI is very low. You can find it on textbook. From you simulation, you can also notice it. Perhaps you can try to run your case on lower TI. me.po likes this.

July 24, 2023, 04:16	did this problem is solved?	#8
vaibhavhiwale New Member Vaibhav Hiwale Join Date: Dec 2021 Location: India Posts: 3 Rep Power: 4	Did this problem is solved? im also facing the same problem

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
caculation of friction factor	Dilip	Main CFD Forum	9	February 23, 2016 06:00
Calculation of Friction Factor & Nusselt Number	mypassion59	FLUENT	1	July 9, 2014 15:31
Getting (Nusselt vs Re) & (friction vs Re) in a ribbed duct	shayan_mv	FLUENT	6	November 14, 2012 04:42
Nusselt numbers for fully developed laminar flow	Danro	Siemens	8	July 19, 2012 08:22

April 28, 2015, 03:23		#3
me.po New Member Mehdi Join Date: Nov 2014 Posts: 4 Rep Power: 12	Thanks, What's your suggestion???

April 28, 2015, 13:42		#7
me.po New Member Mehdi Join Date: Nov 2014 Posts: 4 Rep Power: 12	But that wasn't the cause of the error. when you use periodic conditions, there is no longer any inlet or outlet and flow is fully developed. I checked the problem with finer mesh with very good quality and the problem still remains!!!!!!!!!!!!!!!!!!