Hi, folks!

This is my first time to get here. We are trying to model the air flow in a room using underfloor plenum ventilation. Basically, cool air is ducted into plenum(a whole air body) and get into room through diffuser(with fan or without fan) evenly distributed on the floor. Then the air could be heated up in the room and move upward to ceiling and return through grilles mounted on ceiling. It's like displacement ventilation but one thing different is that there is a higher velocity of the air coming out of diffuser and a better mixing effect(our hypothesis). So we are trying to find a tool which could help us easily identify how the air flow pattern, temperature and radiation distribution are like in the room. We just want to build a 3D box as a room, several other boxes to simulate human being or PC etc, diffuser on the floor, plenum underfloor and a duct outlet on one side of plenum. What we do expect is its' programming ability because we want to add some models into it, like thermal comfort model.

I am not sure about those softwares like Fluent or Pheonics. Is that easy to set up the model in these programs which look like giants to me. For a newbie, how long does it take to handle the program efficiently? Are there some model libraries in those programs that would allow you plug-in in your model? Like diffuser library etc.

Anyway, what do you guys think?

Peter

Peter, it's a very doable problem. I don't know about Fluent or Phoenics, but CFD-ACE+ (www.cfdrc.com) has a concept called 'User Subroutines' that lets you hook in your own models/programs. Most of the commercial codes will have some mechanism for you to hook in your own model. Of more concern is how hard it will be to model the people, equipment, grilles, and the fact that these things are heat sources. ACE can do it easy but others probably can, too.

The learning curve is harder to estimate. Try a demo version of some software. Work through the tutorials and read documentation. Then try first to model a box alone with flow coming in from the bottom at a specified rate and leaving through the ceiling. Then add 'internal blockages' like people and equipment and specify them as heat sources. Then add the circulation subsystem around the box to complete the model of the entire system.

With a good package, working alone, you can be getting results on the first case within a couple weeks. With help, less. Go to the ME department and ask 'how do I model this problem for CFD'.

Dear Peter Zhang

I do not like to say 'this code is better than that code. What I can say is : I am sure that your problem is fairly classical one, so that you problem can be solved by any one of available commercial CFD codes. If you have enough(or basic) knowledge about fluid mechanics and heat transfer, I think it is not giant to you except some (minor?) problem to model fairly complex geometry, for example, human being.

Happen to be, I have had some experience with two commercial codes which you've mentioned, FLUENT and PHOENICS. For FLUENT, it has fairy good models and 'user-friendily' is very good. For me, I prefer Fluent, because I am working in industry and many boss in industries wnat to see the result within very short time. However, if I have enough time, I will use PHOENICS rather than FLUENT. PHOENICS is very good when I want to add specially required model. For example, if one really want to add swirl modified k-e model which was(not 'is') not provided by both codes, one could add his favorite model fairly easily by PHOENICS. Of course, this is relative comparison, especially for my case.

(without the last words, 'for my case', I can not protect the attack of Fluent support group, one of whom, Seung-Eun, is my best friend and he has really lot of knowledge about CFD. Sorry, Dr. Kim, to mention your name.)

Sincerely, Jinwook

(1). It is fine to use 3-D boxes to simulate the room, the human being, and the PC. (2). But in reality, the flow around these objects is quite sensitive to the detailed geometry. So the flow field part is going to be non-realistic. (3). It maybe possible to develope a control volume model where you can focus on the global balance of mass, momentum and energy. In this case, the flow field will have to be modelled, that is relative velocity to the objects. (4). If this is not what you have in mind, then be prepared to dive into the 3-D world of CFD. And each case will be different, whether you add a person, or a PC to the box.

Peter,

FLOVENT from Flomerics is a tailored CFD code specifically designed to model the built environment. I'm not a salesman so I won't try to impress you regarding its useability, stability etc. However, we do have square, circular and swirl diffuser items to name but a few plus a whole range of library items that result in quick drag-drop model building. The learning curve is very short (especially compared to the more collosal general purpose CFD packages that you mentioned) making it an extremely productive tool.

Check out:

www.flovent.com

We've had a lot of experience modelling a whole range of ventilation configurations, including variations on classic displacement ventilation, which we can certainly share.

Good luck!

Robin.