What CFD package to model an atmosphere?
 

Hi.

The company I'm working in is planning to model the atmosphere of Mars with the help of a CFD package. The software will then be used for other applications, mainly in astronautics and thermal analysis.

I'm a completely "newbie" in the computational engineering subject and it's natural that I'm having some difficulty in understanding if a program is capable of doing what I expect it to do.

Basically, it will have to deal with the atmosphere of an entire planet, the wind, the heat inputs from the soil and the Sun and their interaction with a spacecraft. Before modeling the interaction between spacecraft and atmosphere we have to find out which are the best places for landing (this will depend on pressure, temperature variation and wind conditions). This is the tricky part for me. Are the common commercial CFD packages capable of doing this? I've been looking at the websites of ANSYS, CD-ADAPCO and FLOMERICS in a total of 5 software packages!

My questions are: 1. Can anyone tell me which software can model a planet and atmosphere? 2. Which one is the most user-friendly for a non-engineer? 3. I know they are expensive but I can't find any values! What is a price range for the packages I mentioned? 4. Finally, will I need a CAD program to design the planet?

Thanks in advance for your answers/opinions.

Regards, Pedro

Re: What CFD package to model an atmosphere?
 

Hello Pedro,

What you really need for the planetary atmosphere is not an engineering CFD package but more like CFD code from the Scientific side, say either from Metereology or from Astrophysics/Astronomy. I know a few people in Physics/Astrophysics who have been working on modeling planetary atmospheres, usually they use a hybrid approach of stratified layers, where each layer is more like a shallow water equation approximation.

There is Ingersoll A.P. in Caltech (Pasadena) who has been working on Geophysical Fluid Dynamics for many years.

I personnaly know James Cho who has been working on modeling the atmosphere of Jupiter using a spectral method. James Cho is presently in Queen Mary (London, UK).

You should make a search on the ADS:

http://adsabs.harvard.edu/abstract_service.html

on that subject and on Cho James and Ingersoll A.P. and then you might even want to try and contact them or find other people working in that field. I doubt you will find what you need here on this CFD, though you never know...

The modeling of earth atmosphere is only done from day to day, or at most a few days... I am not aware of people modeling the whole atmosphere of Mars for similar (meteorological) purposes, usually people just try to "mimic" and obtain some results in agreement with the astronomical observations. I would suspect NASA/JPL has probably people working on a martian atmospheric model in relation to all their landers/orbiters. You might want to try and contact some of these folks too maybe ?

Good luck.

Re: What CFD package to model an atmosphere?
 

I agree with Patrick's accessment that you don't really want a CFD package.

Climate modeling is a research topic of considerable importance right now. The US government is a big player (take a look at http://search1.ornl.gov/query.html?c...qs=&qc=&pw=100%25&ws=0&la=en&qm=0&st=1&nh=10&lk=1&rf=0&oq=&rq=0& si=0 ). I got that from going to www.ornl.gov and putting 'climate modeling' into the search box in the upper right hand corner.

Mostly the resources required are enormous. The target is simulation of global climate change over decades or centuries.

A step down from that is weather forcasting. Here I think they get by with looking at sections (like one continent) of the earth's surface and for only a few days at a time.

Both these undertakings are horribly expensive and time-consuming.

You might look for a much simplified approach to get results for a 'reasonable' budget in an 'acceptable' time.

Sorry to give such a fuzzy and discouraging answer.

Re: What CFD package to model an atmosphere?
 

Thank you for the quick replies!

In case I can get some data for the Martian atmosphere (I found a website which provides it: http://www-mars.lmd.jussieu.fr/ ) then I will need a CFD package to model the interaction spacecraft-atmosphere. The company is really interested in acquiring such a package (maybe to use it again for other purposes...) and I wonder which one will be easier to use and not on the expensive side...

What I'm looking for is a light comparison between the main packages: CFX, FLUENT, STAR-CD, STAR-CCM+ and EFDLab. Even if they cannot be used for the atmosphere, they will be used for the spacecraft entry analysis.

Pedro

Re: What CFD package to model an atmosphere?
 

If you want to make a actual prediction of the atmosphere (the interaction of the spacecraft with the atmosphere should probably be done with a separate code that takes the atmospheric conditions as initial/boundary conditions for the spacecraft) then you will need a full atmospheric model + data assimilation code. If you're worried about the cost of the software then you're worrying about the least expensive part of the problem (you'll either need to purchase a supercomputer or rent the resourses from another company). If you purchase (most atmospheric codes are free for academic purposes) then you are talking about < £70,000 (and the licence will only be sold if it is in their interest - on going support costs etc) for the code while the cost of the computer will run it on will be in the 10's of Millions!

Basic rule of thumb with the big supercomputer companies is that if you need to ask the cost you can't afford it!

Re: What CFD package to model an atmosphere?
 

From the comments I guess you already know how challenging the atmosphere project is, maybe get a consultant to do it for you, most of the codes have a consultancy wing so you should be able to tie up on the software purchase for the other applications.

Speaking of which I think any of the codes you are talking about will probably do the Job, the "big three" will definitely do the job I have used Fluent, STAR-CCM+ and Phoenics (not one of the big three), Fluent is in a bit of a state of flux at the moment being integrated into workbench, and I have heard they are killing people on license costs at the moment (ANSYS trying to get their money back I think) although I personally haven't seen it being a star user currently, I can't really speak for anyone else but STAR-CCM+ really is great to use and does everything I need but you should get evaluations of all of them and choose yourself!

Re: What CFD package to model an atmosphere?
 

Your choice should also depend on type of modelling you are considering to do. I personally have worked with two codes Fluent and CFX. If you want to do LES modelling then probably CFX will not be a good option because it is fully implicit code. Of these two giants i personally would recommend Fluent even though CFX has got more sophisticated features in it.

yo man

Re: What CFD package to model an atmosphere?
 

I think you'll find none of the commercial engineering codes can do the atmospheric problem - basically certain balance constraints must be satisfied at lead order before detailed fluctuations can be determined; i.e. hydrostatic balance (atmospheres are highly stratified) and quasi-geostrophy (Rossby, Kelvin, ... waves need to propagate correctly). There are also numerical issues at the poles! Since the commercial (engineering) CFD codes were not constructed for this it's unlikely that they can pass the simplest test case for atmospheric dynamics.

On top of this there are also issues with providing initial conditions as well as microphysical processes (e.g. clouds - the different states of water etc in the air).