Atmospheric module : boundaries and meteo file

Questions and remarks about code_saturne usage
Forum rules
Please read the forum usage recommendations before posting.
Post Reply
SimBu
Posts: 30
Joined: Tue Sep 13, 2022 4:13 pm

Atmospheric module : boundaries and meteo file

Post by SimBu »

Dear saturners,

I have dealt with the atmospheric for only few weeks, so I warn you, I'm still a beginner.

I am trying to run a simulation where wind blows over a mountain. The case is pretty easy to understand :
  • I want a transcient simulation,
  • Wind is blowing from the east,
  • North and south are symmetries
  • The floor is divided in 2 : ground and mountain
  • Top is set to inlet, meteorological profile from data (such as advocated in the user guide)
  • The domain is : 6 km of length (3 before the ountain and 3 after). - I know it is a little bit to small but I will resize it soon. -
  • For now, it is dry atmosphere but I want eventually humid atmosphere
I initially carried out the study in two dimensions, before launching the calculation in 3D and nothing was needed to be reported.
However in 3D, my calculation diverges regardless of the efforts made to stabilize the model. I checked several times my mesh which is ok (salome mesh quality report linked below).

** With automatic meteo file generated, I tried to change the inlet / outlet :
The simulation crashes when :

1) east: inlet, meteorological profile from data, Automatic inlet/outlet nature from data
west: inlet, meteorological profile from data, Automatic inlet/outlet nature from data

2) east: inlet, meteorological profile from data,
west: outlet, meteorological profile from data

However, it runs pretty well when

3) east: inlet, meteorological profile from data
west: inlet, meteorological profile from data

I do not understand why... The user guide advocates to set the (1) but it is the one wich crash the earlier....

** With User meteo file provided (the parameters are set really close to the automatic profile)
the simulation crash after 15 iterations or less, whatever my efforts are. It seems to be really unstable.
I have attached two screen : the first one is the velocity with automatic meteo file, and boundaries case 3). With the exact same parameters, the other screen is the User defined meteo file.

So I have few questions :
a) Am I doing all wrong with the inlet boundaries ? and if so, what is advocated, why is it crashing faster withe userguide recomandations ?
b) What can I do to stabilize a simulation without structured mesh (I have already tried inflations, forcing upwind scheme and blending factor).
c) Is my meteo file wrong ?
d) Why does the custom meteo file seems to make the simulation unstable regardless of the setup ?
e) Is it normal that my simulation is much much slower with meteo userfile ?

Thank you very much for the help, it will be appreciated in any case

Please find attached the meteo file, the two screens, the mesh quality report from salome, and the two similar xml.

PS : Don't mind the time setting, I have tried everything, and the constant is currently the more robust, but I tried with time variative. I tried also with Simplec, but inner iteration seems to work well too...
Attachments
mountain case.rar
(426.77 KiB) Downloaded 46 times
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: Atmospheric module : boundaries and meteo file

Post by Yvan Fournier »

Hello,

I mentioned your post to the meteo team.

As far as I known, in most cases, with the newer versions of the code, the "Idealized large scale meteorological data" option should be easier to use, and "meteo" files only used for more advanced cases.

But regarding the other questions, I'll send them a reminder. I am not familiar enough with that module myself to give a proper answer.

Regards,

Yvan
SimBu
Posts: 30
Joined: Tue Sep 13, 2022 4:13 pm

Re: Atmospheric module : boundaries and meteo file

Post by SimBu »

Hi,
Thanks a lot Yvan, your help is very welcome (such as the one of the other dev :) ).

I agree, the "Idealized large scale meteorological data" is more confortable, nevertheless I can't have any control on humidity with that way....

Given that my questions will may be asked to the meteo module's dev, I have plenty of them :

- From what height is it interessting to use atmospheric module (what is the appropriate height of the domain)

- Is the roughness of the wall very important in the model? Can it be the origin of the crash ?

- In the meteo file, k end ε are detailled, is k-ε the only relevant turbulence model or k- ω SST would be useful (in that case, is the meteo file supported?)

- How is the idealized meteo file generated ? Is it from real meteo data ? Or only why a parametric equation including the latitude and longitude ?

- I know that in the reference meteo file, the number of droplet is in (n/cm3), but what is the unit of the droplet flux ? Is it n/cm²/s ? What is the unit of the ym_water (TotWater), it is not written in the user guide, is it ? In which parameters can I detect vapor condensation ? It seems to be LiqWater, hawever droplet do not belong to this parameter ?

- I have very few knowledge about meteorologie, but i understand the nebulo frac as modeling the fraction of the sky covered by clouds, am I wright ? And what about nebulo diag ?

Thanks you,
Post Reply