Best CFD program for analysing windflow in a city
 

Hello,

Basically for my final year design project (architecture) im lookin to analyse windflow around my site which consists of high rise and low rise buildings (modelled in autocad).Im looking to use the resultant flow pathlines to help generate forms for my design. I've just had a quick look at flowizard and it seems to do a nice enough job... but i can't seem to export the actual flow pathlines so i can begin to play around with them in autocad/3ds max. I was wondering if there was another CFD program which would best suit my needs?

cheers

Re: Best CFD program for analysing windflow in a c
 

you can use CFDesign which will suit your applications because it has more options for your applications

Re: Best CFD program for analysing windflow in a c
 

Nah, CFDesign isn't up too it, tet only, no LES, limited number of boundary layers, etc etc. Go for one of the big 3

Re: Best CFD program for analysing windflow in a c
 

LES on city scale... wow, I'd like to see that.

For ease of grid generation (depending on complexity) I would also look into codes with capability to use Cartesian grids.

Re: Best CFD program for analysing windflow in a c
 

LES on city scale... wow, I'd like to see that.

I think when LES was first used (in 60s), it was for weather. So I think we have already seen that.

But I do agree with you, that making the mesh sizes we make with for LES cases, it would be dream to do it.

And to make things little more clearer Ben only said that CFDesign is not upto because it does not have LES model, he did not say that you can do LES for city scale. Its just you understood him that way.

Re: Best CFD program for analysing windflow in a c
 

which are the big 3?

Re: Best CFD program for analysing windflow in a c
 

It could be a matter of opinion, but I guess that would be Fluent, Star-CD and CFX, roughly in that order.

Re: Best CFD program for analysing windflow in a c
 

As a general piece of advice on selecting a product that you can use for your project, where the CFD requirement sounds to me to be relatively superficial - use a package that you can easily get help with. To do this CFD project correctly is actually a huge task, and if I read your initial post correctly, it is not really the primary focus of your project. So try to find a friend who knows how to do CFD, and use what he uses.

Re: Best CFD program for analysing windflow in a c
 

www.winsim.com

www.aria.fr (used by U.S. government for pollution studies, you can have meteorologic entries)

www.simulent.com

http://www.watcfd.com/

CFD for the built environment
 

One of my friends at a government agency sent me a link to your post.

First, the disclaimer: I work for CPP Wind Engineering and handle their CAE applications. CPP also has two atmospheric boundary layer wind tunnels that are used regularly for atmospheric dispersion and cladding load studies. We do a lot of comparison of CFD results to those obtained in the ABL wind tunnels.

Next, how to move information around. AutoCAD is not CAE compliant. What that means is you can't import and export from AutoCAD to the major CAE/CFD vendors ANSYS(CFX,Fluent,ICEMCFD), Star (Cd, CCM+), Numeca, etc. You probably should look at a real 3-D modeling system like Solid-Edge, Solid-Works, Catia, or UG.

If you are trying to model real atmospheric boundary layer flows taking into account turbulent fluctuations and length scales unless you have 10's of computer processors you are not going to get results that compare well to those obtained in an ABL wind tunnel. (I know a major company is modeling a few buildings under a government contract and they are using over 300 million cells and expect to get marginal results).

If you just want to get a very basic understanding of the flow patten around your buildings (no loading, no turbulence, no unsteady flows, no pressure coefficients) you can get something passable using most commercial CFD codes and roughly 3-6 million cells depending on your domain size. This can be run on a small compute cluster. You need to work in model scale to ensure Y+ is somewhat reasonable.

You will probably need a fluid dynamicist to help you decide if the results are more than just colorized fluid dynamics. Most commercial CFD codes are rather forgiving and will produce a color picture for ridiculously configured cases. You need to make sure your results, pretty as they may be, actually have some physical validity.

FYI: Some people are looking at Meso-scale systems, but they cannot read any detailed geometry. This area is left to the pros with big budgets ($1 million +)

Re: Best CFD program for analysing windflow in a c
 

What some people call LES should really be XLES, or do you think the "LES" computations they did on weather in the 60s have much to do with what we mean by LES today? I am sure the meaning of this generic term has changed quite a bit. I have heard of people perturbing their flow model with a random generator and calling it "turbulence". Those were the early days of CFD...

Re: Best CFD program for analysing windflow in a c
 

I think I already said this:

"But I do agree with you, that making the mesh sizes we make with for LES cases, it would be dream to do it."

Re: Best CFD program for analysing windflow in a c
 

uh, so... back to my point: I'd like to see that happen, one day. You want to keep going around in circles? ;D

Re: Best CFD program for analysing windflow in a c
 

Nobody wants to go around in circle, I was just pointing you out that you did not read my post properly, I already agreed with what you said about LES for city scale. We all want to see it happen one day and we all know its not possible with today's resources. This is univerally accepted truth. Its just you want to be mentioning it around, I feel it is like someone keep telling others that sun rises (a universal truth), we all know this.