Computational Fluid Dynamics (CFD) workbench using OpenFOAM

Post by **oliveroxtoby** » Sat Sep 28, 2019 8:54 pm

Crossleyuk wrote: ↑Tue Sep 24, 2019 9:31 pm Okat Oliver, thanks for the new info. I hadn't noticed the additional surface mesh component. Here it is and it still fails soon after commencing the CfdAnalysis Run after a successful Write selection.

Michael

12in VXA CFD Axial Diff 5000 Iter 25-07-19 - Rev P2.2 No Surface Mesh.FCStd

I've made some tweaks to the workbench. Attached is a version that should work after you have updated your workbench. I changed a few things in the file, not all of which will actually have been necessary. Hope that gets you back on track.

Crossleyuk · Post by **Crossleyuk** » Fri Oct 04, 2019 9:07 am

Thanks Syres and thschrader

I did discover that about the Multi Region Mesh cube, it was a selected units problem. But I cannot use it at all in 0.81 with the new Workbench. I'll now move on with Olivers latest suggesten and changes. Fingers firmy crossed, I love what you are ALL doing and your open source environment.

Thanks again ALL!!

Michael

Crossleyuk · Post by **Crossleyuk** » Fri Oct 04, 2019 11:00 am

Hi Oliver,

First of all THANK YOU!! Wonderful job and massive success for me. Not only have your changes solved the running problems but the model runs even better that before in FC 0.17. The residuals are greater than 1000 time lower than previous 0.17 runs.

So impressed and love your open source efforts even more.

P.S.
Still problems with ParaView 5.4 but working via thier Forum on that. Sure our (My) concerns are now behind us.

Michael

Crossleyuk · Post by **Crossleyuk** » Sat Oct 05, 2019 11:08 am

Hi Oliver,

Great progress but I need to gather a better understanding, after a lot of help from you. I have had great results but only if I follow my previous expectation that I had to use absolute pressures.

This is from my current attached model that was fully built in 0.18 and now your revised CfdOF. The two plots are using the gauge pressures in your example and absolute pressure with associated iterations charts.

I now have a task to investigate what I'm being told.

Million thanks again!

Michael

Crossleyuk · Post by **Crossleyuk** » Sat Oct 05, 2019 12:23 pm

Your changes have confirmed that I have a lot of work to do to develop my CFD and boundary conditions understanding. In particular turbulence intensity and eddy length scale. To be clear, my pressure boundaries were:

Absolute Pressure

Control Inlet = 1e5 Pa

Radial Inlet = 1e5 Pa

Outlet (Open) = 98 kPa (Negative 6in water column gauge fixed by a fan)

Gauge Pressure

Control Inlet = 0 Pa

Radial Inlet = 0 Pa

Outlet (Open) = -1500 Pa (Negative 6in water column gauge fixed by a fan)

I'm sure these are valid options but I need to understand the CFD concepts better.

Thanks yet again, I'm now back in the driving seat.

Michael

Post by **oliveroxtoby** » Sun Oct 06, 2019 7:30 am

Crossleyuk wrote: ↑Fri Oct 04, 2019 11:00 am Hi Oliver,

First of all THANK YOU!! Wonderful job and massive success for me. Not only have your changes solved the running problems but the model runs even better that before in FC 0.17. The residuals are greater than 1000 time lower than previous 0.17 runs.

So impressed and love your open source efforts even more.

P.S.
Still problems with ParaView 5.4 but working via thier Forum on that. Sure our (My) concerns are now behind us.

Michael

Great. Thanks for the feedback.

Post by **oliveroxtoby** » Sun Oct 06, 2019 7:33 am

Crossleyuk wrote: ↑Sat Oct 05, 2019 11:08 am Hi Oliver,

Great progress but I need to gather a better understanding, after a lot of help from you. I have had great results but only if I follow my previous expectation that I had to use absolute pressures.

This is from my current attached model that was fully built in 0.18 and now your revised CfdOF. The two plots are using the gauge pressures in your example and absolute pressure with associated iterations charts.

I now have a task to investigate what I'm being told.

Million thanks again!

Michael

For an *incompressible* simulation (simpleFoam), gauge and absolute pressures can be used interchangeably, since only the gradient of pressure appears in the equation. Gauge is generally a slightly better choice to avoid any floating point roundoff errors.

So, you should not get significantly different results. Are you sure you made the change everywhere, i.e. boundary conditions and initial conditions?

Crossleyuk · Post by **Crossleyuk** » Sun Oct 06, 2019 10:28 am

Thnanks for the feedback. I couldn't understand why gauge or absolute should make such differences as you say pressure difference must be the driver in incompressible flow. I'm now repeating with increased numbers of iterations but will check out your initial conditions suggestion next. Quick question, I see that I can change the initial conditions in "InitialiseFields" and they are set to 0 Pa for my gauge and absolute cases. It also seems to apply to all bondaries, should I be able to set may chosen starting pressure independant for inputs and output (open)? Love these learning conversations!

Crossleyuk · Post by **Crossleyuk** » Thu Oct 10, 2019 11:08 am

OS: Windows 10
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.18.16117 (Git)
Build type: Release
Branch: releases/FreeCAD-0-18
Hash: dbb4cc6415bac848a294f03b80f65e888d531742
Python version: 3.6.6
Qt version: 5.6.2
Coin version: 4.0.0a
OCC version: 7.3.0
Locale: English/UnitedStates (en_US)

Tried to do this in a pdf but too large to post in forum so here are the words. Sorry complicated but can't explain.

Let me try to put a clear story together that I have copied to FreeCAD (CdfOF) and ParaView forums.
I have recently had fantastic help and support from the FreeCAD Forum that has solved one of my recent problems and has resulted in modifications to the FreeCAD Workbench CfdOF.
I can now generate the datasets I require from my CFD model, but I’m now left with major problems interpreting them in ParaView.
Starting with my FreeCAD CfdOF model:
10-10-19 FreeCAD Labels & Attributes.PNG

And from PV 5.4.1 launched from FreeCAD and CfdOF:
10-10-19 FreeCAD Glyph Labels & Attributes.PNG

I feel that this is confirming that FreeCAD CfdOF is providing me with a valid model result.
My current problems centre on my understanding of ParaView operation and how it is integrating with my FreeCAD CFD modeling results.
A run from my copy of PV 5.4.1 directly run from a Windows 10 icon provided me with:
10-10-19 PV 5.4.1 Properties.PNG

And also, from PV 5.7.0 from my Windows 10 icon.
10-10-19 PV 5.7.0 Properties.PNG

So similar results but my observation is that the Properties panel only includes model components and not velocity, pressure etc. properties. And no orientation array options.
10-10-19 PV 5.7.0 Properties Glyph.PNG

I must also repeat that only PV 5.7.0 allows me to open a file to edit without launching it from FreeCAD CfdOF and can’t save anything then.

ParaView Copy details (Next Post)
10-10-19 PV 5.4.1 About.PNG
10-10-19 PV 5.7.0 About.PNG
10-10-19 PV 5.4.1 FreeCAD CfdOF About.PNG

Many thanks again.

Michael