Has anyone had any experience using FDS as a general CFD tool for modelling HVAC airflows?

Specifically, we want to use it for modelling air temperatures in the range 8ºC to 28ºC, and air velocities in the range of 0.05 m/s to 2 m/s.

Would Fluent or similar be a better (but much more expensive!) option?

What would be the limitations of using FDS rather than something like Fluent?

I would love to hear from anyone with comments regarding this, whether they be recommendations, cautions or good experiences.

Thanks,

Doug

I have not used FDS, but have read through the docomentation in some detail (and met one of the main developers and current custodian of the code at a workshop). The code is well-documented and I believe is fairly well-supported (given it is free and mostly a one-person support). The short-comings would be the following (which might not apply to your intended use):

1) The grid used is strictly Cartesian. If you have curved surfaces that you want to resolve, you will need to live with using stair-stepped grids and the associated error.

2) The grid data is entered through a text file, basically specifying the x,y,z min/max for each block of the grid. You can have internal obstacles. If your case is complex (lots of blocks or obstacles), this grid input format could be an issue. You might check if any commercial software (not free) is available to create the FDS grid data format based on some existing grid formats you have, or from CAD files.

3) Recent versions of FDS are parallelized, but the pressure solver used by the code is not "truely" parallelized, meaning that some error is introduced in mass conservation as the number of processors increases.

You can try ICEPAK, it is a Fluent product, but solves the incompressible Navier Stokes equations and the Energy Equation. Has fast convergenc, Automatic meshing and above all is very straight forward to use.

To be honest with you, I have just downloaded FDS, and I am struggling to understand it. The data input is a real issue that you have to think about very seriously, the code might have been validated, I have no comment about that but without a GUI it is just a nightmare. If you are looking to solve industrial type applications with real geometries you have to think about ICEPAK, putting the two together I would stick to ICEPAK, I was asked to do some heat transfer analysis and I have never used ICEPAK before, all I can tell you is that by the end of the day I was having my analysis almost finished. A GUI is a must in any real world analysis and for the range of velocities and temperatures you have mentioned (Incompressible) I would recommend using ICEPAK

I think you might mean Airpak.

No Harry, my comments are based on my experience with ICEPAK. If your experience with AIRPAK has led you to believe that I was talking about AIRPAK, well there is no problem with that both are products of FLUENT and they have definitly used the same core technology in both. Glad to know this fact

Only in that Icepak is designed for electronic thermal management, Airpak for prediciton of the built environment. True though that both are codeveloped with ICEM CFD using the ICEM CFD auto hexa and tetra tools as base and as such there ain't much separating them!

Hi thanks for all your input! It is much appreciated.

My specific problems are to do with air-conditioning systems for rooms. I want to model the performance of displacement air-conditioning systems and more conventional aspirating types.

I have had some small experience of fluent, and I know it is excellent. I would like to get a feel for the "trustworthiness" of results from FDS given the low air velocities and small temperature gradients I am dealing with compared to fire and smoke simulation.

Doug.

Your range of velocities(Low Mach Number Flows) and temperatures is in the incompressible range. I guess you have to go through the documentation of FDS (FDS Technical Reference Guide, Chapter 3) and check their model to see if it is adequate for your HVAC model. My educated guess is that you can use FDS for your application, provided that you do not use the combustion model. If the budget is very tight and you are looking for a free software why not check Openfoam, it is an open source code similar to FLUENT which you are familiar with. There are questions that you and only you can answer, for instance what is the degree of your real expertise in Fluid Mechanics and CFD? (there is no support engineer on the other side of the phone to answer your questions) what are the time constraints on your analysis (Using FDS you will have to input the geometry and other data manually) etc.., and if you have found a free programme to help with this task I hope your feed back will help us too. I wish you luck. AIRPAK

Hii Doug, reccomend you look at CFX as well. This is cheaper than Fluent and can be as complex as you require it to be. We used it a lot looking at air flows within rooms / Atria - mesh's are simple to setup and their support at CFX are great - which is an important consideration when first learniung to use a CFD code.

Fluent also can do all of the above, but it may be more expensive, again depending upon the options you choose. Get the code vendors to come around and demonstrate for you ! There is even Flowvent - a code which is basic in terms of all round CFD but which is predominantly used by HVAC eng.'s

Dave

Hi Diug Crawford

I would like to see your input file of FDS and some detail of the problem. FDS is extensivly validated tool and I am using it from last 8 years. I see no reason for "trustworthiness" of FDS. Infact its not a general purpose tool but it does the weakly compressible type of computation quickly and with great accuracy. You can email me at pa1.sharma@gmail.com.

Pavan Sharma