topic Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga in CFD Forum

Gravity and Heat Transfer not working with Large External Flow and Mixing Gases

Anonymous — Thu, 20 Oct 2016 20:31:20 GMT

I am doing an analysis of heavier than air gas concentrations from several vent sources outside of a large building (~1E6 sq feet). I've got the model/simulation working if I include everything EXCEPT heat transfer and gravity, but as soon as these variables join the fray the system comes crashing down. I am totally stumped.

On the attached .cfz, the (reasonable and beautiful) results at iteration 200 show the model with prevailing winds, and flow with scalar quantities from the vents in question.

I have tried several Advection schemes (1, 2, and 5) while trying to get temperature and gravity incorporated.

What am I missing? I really need gravity to be included as the gases in question are heavier than air and I need to look at their potential concentrations at ground level. I think temperature is less crucial but I can't figure how to have gravity without Heat Transfer.

Having trouble getting the share file to upload. I've included two screenshots to show the contrast - the only difference is that I turn on Heat Transfer (and therefore gravity) in "Capture". I have gravity pointed down (0,0,-1).

Thanks for any help!

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Fri, 21 Oct 2016 14:55:39 GMT

If you need to see the cfz, I can place it in a dropbox folder (the file is too large to place here).

The only thing I change (that causes the solver to choke and the results to go haywire) is shown in the attached image, Capture3.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Mon, 24 Oct 2016 12:46:56 GMT

I'm really stuck on this and the need to figure it out is increasingly urgent. Let me condense the issue to some basic questions to try and draw out a few responses and start this conversation:

1 - Is there a way to get the simulation to recognize gravity WITHOUT asking the simulation to worry about Heat Transfer? I need to see where my heavier-than-air exhaust is accumulating.

2 - I'd LIKE to include temperatures and heat transfer if possible. What would make the simulation freak out (1E6+ft/s wind speeds??) just by adding temperature, heat transfer, and gravity? Are there basic things that should be included when assigning temperature et al that I could be missing? Refer to screenshots in previous posts to see the boundary and initial conditions I placed for temp.

Thanks,

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Mon, 24 Oct 2016 14:53:19 GMT

The simulation has air flow out of the TOP of one of the two zero pressure boundaries on the external volume and IN to the BOTTOM of that boundary. How is this possible? I don't have any temperatures extreme enough to justify this kind of convection. See attached image.

For clarity: the boundary condition on the SOUTH and EAST faces of the external volume are a 2ft/s wind at 20° C and a scalar value of zero (air). The boundary conditions on the NORTH and WEST faces are just zero psi. The bottom face is a slip/symmetry boundary.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Tue, 25 Oct 2016 14:00:56 GMT

Hi Dean,

I think you have to share a cfz file, without it may be impossible to help...
some things to think about:

- is the external flow clearly in one direction defined?
- is the domain large enough?
- are there areas of (re-)entering fluid without having a temperature or scalar value defined (this one should cause your divergencies)
- maybe try first without scalar, just temperature. You can stabilize the scalar solution with a higher diffusion ratio.

Marco

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Tue, 25 Oct 2016 14:12:38 GMT

Marco,

Thanks for the response.

I have the external flow coming form the southern and eastern planes of the external volume. So, there are two directions defined here. The simulation had no problems with this before I added temp and gravity. You think this could cause the instability?

The external volume is 5 times the building height and extends far beyond the "ground" that the building sits upon. What is "big enough"?

All entering fluid values have temperatures and scalars associated with them. You think the simulation would work better if I removed these added BC's to some of the fluid entry points?

I have not tried doing the simulation with just scalar and not temperature (although it does work without temp and just scalar). The diffusion ratio is well above the Autodesk recommended value.

I don't have a place to share the cfz from and it is too big to place here.

Thanks again for responding!

- Dean

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Wed, 26 Oct 2016 06:57:00 GMT

usually these instabilities are caused by little regions of re-entering flow (at dedicated outlets), where then scalar and temperature are "undefined". Thats why the external flow should have a clear direction and the domain should be big enough.
Everything that is inlet should maintain with scalar and temperature!

Also a high aspect ratio of the finite elements could cause this, you could search for "stream function" here or in the help system, which can help you to identify if this is the reason.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Wed, 26 Oct 2016 18:05:42 GMT

All of my inlets are clearly defined (with velocity, temperature, and scalar quantities).

The only "outlets" are the two zero pressure boundaries.

Refer to screenshots from my first post to see the Design Study Bar.

I am keeping my element count between 900,000 and 1.5 million.

Is there a way to have gravity affect my densities while ignoring temperature? I get reasonable results until I add those two variables. Since gravity is more important to my final results, I am stuck on the hope that temperature is the problem and that I can cut it out from the solution to still get the data I want.

Very stumped.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Wed, 26 Oct 2016 18:24:13 GMT

Even if I compress the cfz it is 128,325 KB. If someone has a dropbox, I can toss it in so that another set of eyes can explore this problem.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Thu, 27 Oct 2016 06:28:07 GMT

gravity is always related to heat transfer. I suspect there is more a numerical than a physical issue in this model. Looking at your screen shots, 1 million elements could be too few, though its hard to say...

Did you start with scalar/temperature from scratch or did you use your flow results as initial condition for a restart run? The latter would be preferred...

There are several ways to share big files, out in the web. You can maybe delete the mesh settings or see if there are any big files in the design study folder to make sure to get a reasonable file size.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Jon.Wilde — Thu, 27 Oct 2016 07:11:41 GMT

Taking a step back for a minute and thinking about your inlets. What happens where the eastern and southern inlets meet at an edge?

Do the nodes here both have the same conditions or, if the flow is normal to the surfaces, do they have opposing conditions?

Might it be better to rotate the domain and have a single inlet?

Not sure on the gravity comment - it really depends how the air is set up. Did you change the material so that the density changes with scalar or temperature?

I suspect Marco might be spot on regarding your mesh too.

If you have Dropbox, you could simply share a link to download it from where it is located?

Thanks 🙂

Jon

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Thu, 27 Oct 2016 12:01:46 GMT

File is located here: https://www.dropbox.com/s/8y8xs197jcgfyax/Small%20Capex_Ozone_2.cfz?dl=0

I did solve for flow first and then add the other conditions. I could try turning up the number of elements.

Jon,

Great (and, come to think of it, obvious) suggestion. I was trying to simulate a NW wind with flow conditions from the S and E, but there is an extra high wind gradient where they meet (at the corner). I'll try rotating the external volume 45° to see what happens.

The most important part of this analysis is to identify where gases being vented outside the building are concentrating. The only appreciable difference between air and the venting gas is density (the vented gas has a higher density - the risk is if it pools somewhere on the ground). Currently, I have the density of air varying according to the scalar, not according to temperature. Could that be the issue? It certainly isn't something I've considered yet.

Thanks so much to both of you for your feedback.

Cheers,

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Thu, 27 Oct 2016 12:53:58 GMT

as Jon mentioned, the external flow conditions need to be sorted out.

Also the mesh has to be refined and the external air domain must be larger and must not go below the building (if it is created in CFD where it has to cover the whole building, try to model it in Revit)

Regarding the mesh refinement: the extremely small inlets you have inside your model enforce a very small boundary layer mesh (mesh enhencement). In combination with the large overall mesh, we get high aspect ratio elements. you could try to use "surface refinement" mode as well...

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Mon, 31 Oct 2016 12:27:40 GMT

It looks like the "simulated" NW wind was part of the problem. I created my external volume in Revit instead of using SimCFD so that it would not be below the level of the ground and rotated it 45° so that I could create a wind tunnel. Most of the issues seem to have resolved.

HOWEVER, there is one final concern: there are areas where the venting gas (scalar 1) seems to be appearing out of thin air - specifically over the north central part of the building.

I have updated the share file (https://www.dropbox.com/s/8y8xs197jcgfyax/Small%20Capex_Ozone_2.cfz?dl=0).

Any thoughts as to why this is happening and how to fix it would be much appreciated. THANKS!

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Jon.Wilde — Mon, 31 Oct 2016 14:52:07 GMT

Try recessing outlets into all the pipes that you have, this will help to capture the flow as it exits far better than just a flat surface. Most are looking great now 🙂

Can you share an image of where you are seeing Scalar 1? I see that you only have fractional values entered.

My guess is that the mesh here is not fine enough to capture the scalar mixing.

Thanks,

Jon

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Tue, 01 Nov 2016 22:24:00 GMT

Jon,

I've recessed the outlets. My solver keeps crashing now, though. Can't get through more than 80 iterations (even when removing variables like temperature, scalar gradient, etc so that the simulator only has to solve for flow for the first pass).

So, what I am concerned about are concentrations of the gas I have set to scalar 1 at or higher than 0.00000007 (7E-8 or 70 ppb) within 6 ft of ground level. The gas is venting at 200 ppm (thus the fractional values). How fine should my mesh be to capture this?

Thanks, Jon!

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Tue, 01 Nov 2016 22:24:52 GMT

Should mention that yes, I have tried doing (and not doing) mesh regions around each of my outlets.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Jon.Wilde — Wed, 02 Nov 2016 07:53:00 GMT

You would need to test a mesh sensitivity study to be sure that your mesh is fine enough, there isn't really a way to be specific about the perfect size I'm afraid.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Fri, 04 Nov 2016 12:13:35 GMT

The simulation is still struggling to resolve. Since I cannot get it to converge, I cannot do a mesh sensitivity test. The most recent version is in the dropbox.

I currently have mesh regions around each of the vents.

Re: Gravity and Heat Transfer not working with Large External Flow and Mixing Ga

Anonymous — Mon, 07 Nov 2016 08:16:31 GMT

Hi dean,

I have not the time to look into this deeper, but the fact you get "appearing" scalar out of nothing and solution issues seems to show all of this is related to high aspect ratio elements, as I wrote before. Please check that with "stream function", I did it once and found these areas, where you said scalar appears...

And finally the "surface refinement" method could be one solution for this or a manual mesh refinement in these areas.

Marco