Straight Pipe with imported velocity profile. Pressure gradient?
 

Hello,
I have simulated a straight pipeline (2d axissymetric, k-w sst with mesh y+ = 1.05) to obtain a fully developed velocity profile
.http://www.cfd-online.com/Forums/att...1&d=1473074988
After importing profile to the inlet of the pipe, static pressure gradient can be observed at the inlet.
http://www.cfd-online.com/Forums/att...1&d=1473074988
and on XY plot (red)
http://www.cfd-online.com/Forums/att...1&d=1473074988
.
Why fluent assumes constance static pressure profile at outlet instead of consant value of Total pressure at outlet?
Static pressure in pipe should be an inversed SQRT function of velocity (Bernoulli).
Is it possible to set Constant total pressure at outlet to avoid pressure gradients at inlet?

Hi,

I would like to help but I have trouble understanding the problem. Can you clarify a bit what you did exactly? I understand you imported a velocity profile as a boundary condition and now you do not want the respective pressure gradient invoked by it? Or am I misunderstanding?

Some more comments; what direction are you plotting in, radial I assume? the velocity profile does not look fully developed; a pressure drop along the pipe can be seen, which seems reasonable; normal to the walls you should not be able to find a pressure gradient, of course. Why should total pressure be constant?

Cheers, Mo

Thanks for respond.
That's right. I am modelling a pipe in 2d axisymetric space, that's why you can only half of the velocity profile at outlet. I imported a velocity profile to the inlet b.c.
Plot is indeed in radial direction, where 0 correspond to r=0 and 25mm is max radius. This is a fully developed velocity profile with Re=3*10e05. It indeed looks different than a profile generated with Power Law.

The problem is that the static pressure at inlet should have the same profile (const at traverse) as the static pressure at outlet, keeping in mind some pressure loss.
Because I simulate a piece of a pipe, both velocity and pressure profile should be identical at inlet and outlet, so the results can be trustworthy.

Oh, okay! So you are trying to get radially constant pressure as well at the inlet. Get it, I understood the contrary, so I was confused because radial pressure distributions usally only occur within the entrance/develloping region and actually not even there for Re=300k..

Well, without knowing much detail about your set-up, in particular the way you wrote the inlet BC, I would say it looks like the BC is defined to do so..

How did you import the profile, what type of BC is it? A velocity-inlet? That should basically provide what you are seeking to do: velocity-inlet should give a constant static pressure and adjust the total pressure according to the velocity profile you defined.

It looks the other way around here.. It seems to set constant total pressure.. Did you set it up as a pressure-inlet?

Mo

Ps: Re=300k, v~5.5m/s, D~0.05m.. are you sure about that?

I am sure about the numbers. I set Velocity profile with imported profile and constance pressure 200kPa as I am initializing from pressure inlet. Indeed Dynamic pressure profile looks similar at outlet and inlet, because velocity proffile doesn't change much. However I can not obtain constant static pressure across diameter from beginning up to the end. It changes in near wall region.
As far as I know the pressure you set in boundary conditions is the static one. I tought that it may be caused by instant expansion after inlet, because right after inlet there is a first 10 um triangle expansion ( first of the roughness shape,roughness is in a shape of a saw).However 10um expansion isn't much considered the pipe radius is 25mm.

I have two other problems related to meshing.
I generated a mesh in Ansys Mesh. 180k elements, mostly quads, but with some triangles. When I check statistics for mesh quality I can see that
min Orthogonal Quality=0,44
max skewness=0,77
However when this mesh is read by fluent it shows min orth qual=0 and max skewness=1. What's more When I use General>scale to change m>mm I see that y-min coordinate is -9.3138e-35 instead of 0. I checked in geom file and the line coordinates are [x,y] [0,0] and [0, 25].
Any solution to such problem? I am using ansys 17.0 student license.

At pressure inlet which you may not use, you set both total and static.. But I still cannot reconstruct, how you define the boundary condition exactly, so I'm in the dark, you have to provide all relevant info. Sorry.. It still looks to me that your static pressure reacts to the velocity profile, interacts with the wall-BC and then equals out downstream. I'd say redefine your BC thoroughly.

That's new info too. What expasion should there be, is your flow compressible now? for that Re your probably using water or some similar liquid. I have the feeling that is not the problem..

Interesting, you shouldn't have been able to obtain convergence then. However, you may have improted a mesh from a file that (it so sometimes happens) is changed according to the precision of that file and may end up not exactly right - it needs to be repaired. You can do that in fluent, it's easy, check tui in mesh/repair-improve. Usually that fixes such problems.

Mesh problem was solved with /mesh/repair-improve/repair command.

My boundary conditions are:
Velocity-inlet with velocity magnitude imported profile,2 bar pressure for initizalization, Turbulence instensity 5%, hydraulic diameter 50mm.
Pressure outlet- 2 bar constant value
wall- roughness height 0.
axis- wall at y=0.

Thru calculations static pressure should be constant in a traverse of the pipe at a specific location- x coordinate. Problem is that the calculated static pressure profile at the inlet and then in the pipe is deformed and is not constant for specific x, what you can observe on the picture.

Hi, I can't reproduce your result, just quickly set up a similar thing and radial pressures are constant at every x. Try again, reinitialise and let it converge fully. Maybe you get rid of it. In principle I'd choose the same approach, using a velocity inlet. Sorry that I couldn't help better. Good luck!
Mo