Learning Forum

[psayasith]

Are there more settings available for exporting solution data than just the menu in Fluent Menu -> Export -> Solution data. I found that the data that's being exported doesn't have enough digits in precision to accurately convey the cell size I'm using in my mesh. Furthermore, it appears that the export is incomplete and there are several values within the exported data that are populated with 0s. I'm trying to export pressure and velocity data for input into another program and I found that there are several areas wherein the value for all of these is 0.

[Karthik Remella]

Hello,
Did you launch Fluent as Double Precision? What kind of cell sizes do you have? What are your length scales like?
Karthik

[psayasith]

Yes, I'm currently using Fluent with double precision. My cell sizes are 2.5E-4 M. My shape primarily consists of a pair of rectangular channels that are each 30 cm in length and 5x5 mm on the other faces. There are smaller capillary entrances attached to them and another cylinder with a radius of 2.9 cm and a height of 0.375 cm.

[psayasith]

Do you have any other ideas?

[Rob]

What format are you exporting? Is it surface data?

[psayasith]

I've been exporting ascii data for the whole volume. I'm exporting x,y,z velocity and position, and absolute pressure. I've found that there are large regions in the center of my device where all of these parameters are 0. This is where I expect the parabolic flow to be at its highest.

[Rob]

Try writing a profile. Those are used in Fluent cases too so may have more numerical precision.

[Karthik Remella]

Just to add to 's comments - could you also plot the data in Fluent? Perhaps, a contour plot? - just to verify!
Karthik

[psayasith]

I attached a picture of the contour of pressure along the central axis of my device as well as a section of the output file. Notably, there's a region where the pressure is 0 in the output even though that isn't listed in the device. For reference, 0.018 m in the y is the center of the top channel and 0.005 m is the center point in the x-axis, so some of these missing regions are square in the middle. The long rectangular channels are 5 mm x 5 mm on their faces. I haven't had a chance to try and create a profile yet, I'll try that next.

[Rob]

What does the file that Fluent spits out look like? Just want to check it's the export that's at fault and not the reader on the other side. A text editor ought to open the file.

[psayasith]

Here's the file when opened in Notepad++. I highlighted a row that's also present in the image from OriginLab. It definitely looks like the precision issue isn't from Fluent, but I'm still seeing a 0's in the pressure section which shouldn't be possible.

[Rob]

Not sure. The zero entries in the Fluent table could be solid regions, there is no pressure value there. If you turn on energy & add temperature to the table what do you see?

[psayasith]

I added energy and temperature to the table, they are also zeroed out. I know these regions aren't walls because some of these are right in the middle of my geometry, plus I've noticed that Fluent only outputs my volume and none of the space in the walls.

[Rob]

OK, that's not good but is useful. All cells in Fluent will carry a temperature whereas only fluid cells have a pressure. Can you post the ASCII panel to show exactly what you're exporting? I've not seen (or heard of) this problem so it's odd that you are.

[psayasith]

This is what I normally export. If this is a problem you've never seen before, do you think a reinstall would help?

[psayasith]

I installed Ansys 2021 R2, I was using R1 previously, and I still have the same problem.

[Rob]

Does cell centre help? Fluent stores data on the cell rather than the node.

[Adam Venter]

Hi Psayasith

I have encoutered the same issue here. No resolve yet... please let me know if you have since found a solution to this issue. Thanks

[Aditya Lakhotia]

I am facing the exact same issue while writing solution data to an ASCII file. The solution file shows no issues during post-processing, so it's surprising when zero- values are appearing over some cells in the data, especially for density. What's the matter with this? And does anyone have any alternatives to obtain a clean file (but necessarily in ASCII)

[Rob]

Cell centre or node values? 

[Adam Venter]

I encourtered the issue when extracting cell volume values via a udf - 'C_VOLUME'

[Aditya Lakhotia]

Node values. 

[Rob]

Adam - what number type did you add the C_VOLUME data to? 

Aditya - what about cell values? 

[Aditya Lakhotia]

I specifically require node values, so I haven't tried cell values. I see this as a common occurrence, a lot of other posts relate to the same problem. The problem is with the export of solution data as ASCII from Fluent, so what is being done to address it?

 

[Adam Venter]

Hi Rob, I've tried both real and double.. the issue unfortunately persists.

[Rob]

Why specifically node values? If there's a zero value returned it implies that the node doesn't store that data, or rather that the data isn't mapped to the node. 

[Rob]

Adam - how many figures did you export? Just wondering if you were returning 0.000 rather than 6.3e-9 or similar.