Hi everyone, I am a student studying architectural engineering and im interested in CFD. Could someone tell me how about the application of CFD in building design or research? is it a promising and valuble orentation ? thanks.

Its good for research like CFD of architectural structures, and may be oil rigs. Not much commercial value though.

Well its useful with regards to civil engineering. A lot of stuctures particularly in foreign countries are stiving to get higher/larger. Wind around these structures becomes important. Also bridges, tunnels etc.

CFD is used to my knowledge in two ways in buildings, but not extensively (yet).

1)To help with the heating/air conditioning (HVAC)-especially w.r.t. to comfort studies in a room (ie given some heating/cooling flow in, radiation from the window,... Is someone comfortable sitting at their desk?) It has been used in some very large rooms (auditoriums, indoor sports arenas...) to help determine HVAC needs.

2)Wind forces around major structures. All major structures alter the flow air locally and it interacts with surrounding buildings. CFD can be used to predict the wind forces acting on a building.

Could you please define 'foreign country'?

..enough commerical value for both Flomerics and Fluent to develop Flovent and Airpak respectively. CFX is also often applied to model the built environment.

Ready adoption of CFD into the design practices of the HVAC industry has been somewhat slow due to the glacial adoption of 3D CAD. You'll find CFD applied most to prestigous, high cost or mission critical architectural projects.

There was a man working in the same building as me, and he did a presentation on advancements in aero research in canada, and commented on a massive bridge in ottowa or somwhere (sorry, I cannot remember the city, one of the major ones though). The were using CFD in conjunction with FEA etc to predict the frequencies of the bridge, and the wind around it, so as to prevent natural oscillation (like happened that suspension bridge in america). sorry its not much, but its something you may want to put into google or such.

thank you for your information.Sometimes a little information from an experienced engineer is of great value to a student in school.To realize what CFD engineers are doing is important for me. Sooo,thanks again.

Steve, I don't think that most serious engineers will use CFD to predict Forces on Buildings. This is one are that a lot of research is being done in, and until turbulence modelling can be improved / computer speed increased, this will be a long way off. Forces on Buildings are dynamic, pressures fluctuating, with peak pressures being hard to predict even when the most advanced Turbulence models are employed by the research comunity. At present we still depend upon wind tunnels for these tests, but our time will come !! Bob

We are working with the built environment. Have a look at: www.smhi.se/cfd

I think the limitations of CFD and two-equation turbulence models are well know and have been for some time particlulary for bluff-type bodies and separated flows, however, I don't believe these are valid reasons for not using CFD code for industrial analysis.

The same could be said of experimental errors in a wind tunnel for not using a wind tunnel. When undertaking any kind of analysis provided that you are aware of any limitations of the techniques or models used then it can be used succesfully. Not to mention timescales and costs etc.

As for computational speed the average pc can simulate quite large and transient problems and has been able to for some time in particular external flow around buildings which based on my experience are typically geometrically simple. As most buildings are now modelled using CAD results from a CFD simulation should be attainable by a compotent CFD analyst within 2/3 days of receiving CAD data. It is unlikely that a scale model could be built instrumented and time made available in a wind tunnel within anywhere near the same timescale!

Yes I agree about your time scales, but I did not say to not use CFD for industrial problems. What I was trying to convey (being a CFD modeller at a facility where wind tunnel testing is also conducted), that to obtain the results that are necessary for work within the built enviroment, that CFD is at present not capable of delivering the necessary reliable data.

Turbulence is a big issue in an atmospheric boundary layer. Look at all the work done on the Silsoe Cube project that was presented at CWE 2000 international symposium. The variablility of the results of different turbulence models. The cube was the most basic of shapes. The majority of the runs were steady state. Most areas of work within the built enviroment ideally need to be transient, capturing fluctuations in pressures/forces. just representing the Atmospheric Boundary Layer is a big issue initself. These problems are real.

Now you can use CFD for Architectural purposes, but caution must be used. Fine, look at air flows within buildings, these can not be modelled very easily in any other way. Look at pedestrian comfort levels (but don't expect to include the gust effects into your calculations). CFD at present can be an effective screening tool, but don't go calculating your cladding loads based upon it, unless you're prepared to wack on a large safety factor.

It maysound like I don't like CFD. Far from it. This is my main area of work. But you should be aware of the present limitations of CFD for your particular field of interest. Personnally I can't wait to see the day when CFD replaces wind tunnel, I get so much grief of the wind tunnel guys.

Sorry If I'm stating the obvious.

Bob

Room airflows (comfort, pollutant distribution and such) have been studied for years using CFD, just take a look at the ASHRAE trans. The field is growing and just like any other CFD application, knowledge of CFD fundamentals is required to obtain reliable solutions.

I've seen 1 or 2 projects about flow around building exteriors, but for all I know they may be marketing demos more than real engineering work. Sometimes its hard to know what you are looking at. Still as the cost of computing keeps dropping, I would expect to start seeing this more often - certainly its worth some research time in grad school.

Hi steve, I agree with you totally. In time we will get there. At present there is a great deal of research that can be done. The Wind Engineering society is a great place to look at papers in this field. The problem (or one of them) is the obtaining of fluctuating forces. To do this we need LES type methods. To apply an LES type model to an atmospheric boundary layer, possibly 1 to 2 km in diameter is computationally expensive. Work has been done to try and initialise flow fields with the correct velocity and turbulence profiles, matching the matching the velocity fluctuations so that they satisfy desired spectral densities. This may then enable models to become smaller and less expensive. This is even a problem for regular RANs models, where sufficiently large domains are required to produce the correct incoming velocity and turbulence profiles. All this is a very active area of research.

Bob