CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Pipe flow heat transfer (http://www.cfd-online.com/Forums/fluent/121560-pipe-flow-heat-transfer.html)

 feedq July 31, 2013 10:37

Pipe flow heat transfer

5 Attachment(s)
Hello all,

I am attempting to model a simplified model of a copper pipe transfering heated water through an environment. My goal is to determine the heat loss of the pipe with different fluid and environment temperatures.

My conditions are as follows:
323K fluid temperature, 1.5m/s fluid velocity, water-liquid
15mm pipe radius (quartered cross section to reduce complexity), 2mm thickness, 2m pipe length, copper
293K environment temperature

At first I modeled the pipe within an enclosure, created a coupled wall between the pipe wall and enclosure and turned on the energy equation. K-Epsilon was used for the viscous flow model. Radiation was also enabled.

The results I got were very off. The heat loss along the length of pipe is very eratic (as can be seen from my XY plot of temperature along the pipe's length).

Any help that anyone could give would be greatly appreciated.

Regards,
Q

 feedq July 31, 2013 10:40

1 Attachment(s)
Forgot to post the XY plot

As can be seen there is very odd temperature differences along the length. I also attached a picture of the temperature contour plot of the enclosure and pipe which shows no heat escaping into the enclosure :confused:

 feedq July 31, 2013 14:13

I appreciate I haven't explained the problem very clearly. If anyone requires further information please say.

Any help would be greatly appreciated :)

 macfly July 31, 2013 15:04

Hi,

You can just model the pipe without the enclosure, unless you are interested in the flow features in the enclosure? You could just impose a convection boundary condition a the pipe surface, no enclosure. You have to estimate a heat transfer coefficient at the surface of the pipe, see some heat transfer book for that.

 flotus1 July 31, 2013 15:12

s..t in, s..t out:rolleyes:

The mesh you use violates the basic rules for finite volume meshes.
In addition to being way too coarse in the surrounding region near the pipe, there are no prism layers for the fluid outside the pipe.

It is not clearly visible how many cells represent the radial direction inside the pipe. but it looks like 1 or two, again without prism layers.

 rayan24 July 31, 2013 15:25

I confirm that have been said by Alex, the mesh is not clean.. I recall that the finite volume method is conservative therefore made sure that the mesh is clean

 feedq August 1, 2013 07:47

Thanks for the suggestions guys. I removed the enclosure because as stated before I am only interested in the pipe's fluid temperature drop.

Hope it sorts it out! I'll post results soon :)

 feedq August 1, 2013 09:24

1 Attachment(s)
Sh*t in = sh*t out indeed!

Simplifying the model has sorted it out. The results are far more realistic.

Thank you all for your help :)

 feedq August 1, 2013 09:26

Forgot to clarify..

I simply removed the enclosure, changed the wall thermal settings to convection, used 13.1w/m^2.k for heat transfer coefficient (water-copper-air) and 293K free stream temp. Fixed.

heat transfer in a pipe

hi friends
i am a new user of openfoam and i need to simulate the heat transfer of water in a pipe whit the air out side.
can i use bouyantpimplefoam solver.
or i should use chtmultiregionfoam solver.

i create the geometry with wedge type boundry and toposet rule.