# Links between CFD or MZ and thermal simulations for simulating air and temperature distributions in atria

3 posts / 0 new

Hi,

I am currently starting up on simulating a relatively large atrium in Denmark (app. 55?N) with a glazed roof (sloped towards the sun - clever), open balconies with working areas connected to the atrium, as well as a smaller atrium with smaller openings connecting the two atria together with each other. The main focus in the simulation is first of all the air velocity and temperature distribution in the occupied zone and on the balconies - describing the thermal comfort as well as draft rate. The next focus is on integrating the smaller atria as a heat exchanger in winter time as well as analyzing if this can be used for cooling in summer time, transporting the air from outside to the biggest atrium, via the smaller atrium with little heating of the air. The calculations I want to perform in CFD also regards simulations of the solar gain in the biggest atrium. However simulating radiation in CFD simulations require more computing power for bigger rooms with a lot of surfaces. Any advice on simulating solar gain in CFD? How can this dynamic gain be simulated in CFD? Another cool thing in this atrium would be the heat recovery of the extracted air in the top part of the atrium.

Simulating this in a thermal simulation software requires a lot of luck to obtain the correct (and true) result in the first couple of simulations. However I am thinking of integrating the results from a CFD calculation into the thermal simulation. I am thinking of doing this as an iterative process, integrating the results from CFD when needed in the thermal simulations. This will probably result in at least 4 different simulations over a whole year, giving detailed results (temperature gradient, flow between zones, convective heat-transfer coefficients etc.) which can be used in the thermal simulation for these 4 periods over a year. As mentioned above simulating radiation in CFD is tough, I am thinking of leaving this calculation in the thermal simulation.

Another simpler setup would be the integration of a multi-zone calculation in the thermal simulation giving a more coarse result of the main flow route in the atrium and mixing between balconies and the atrium. However these results could potentially give the same detailed results I am looking for.

Finally has any one worked on an uncertainty analysis of these coupling methods illustrating first of all the uncertainty in the results as well as the sensitivity in the input parameters?

Has anyone done something like this before and what tools were used? I have been given free hands with regards to choice of software so any feedback on this subject will be useful. I'm really looking forward to hear from you :-)

Kind regards

Frederik Vildbrad Winther (FRW)

Offline
Joined: 2011-10-02
Reputation: 0

Hi Frederik,

DesignBuilder CFD would allow you to come up with an EnergyPlus multizone
air flow linked thermal simulation coupled with a steady state CFD solution.
This has recently been released and is a cost-effective way of linking the
2. If you used DesignBuilder to develop the energy model/Steady State CFD
solution carried out and then use their model export options to give you the
model upon which to then import it into a more complex transient CFD tool
should it be necessary. To some extent you could argue whether or not you
need the transient analysis as the dynamic simulations should give you the
temporal solution for the airflow and then you could look at what happens
for say peak or mean flow conditions perhaps. But I?d suggest that it might
well be worth it from what you are describing as even a well set up thermal
multizone model would do most of what you describe, but the thermal
stratification issues are going to be a sticking point for most simulation
tools, both Dynamic or CFD to some extent.

If you really wanted to see the time effects of the Solar Gain on airflows,
etc, there are many CFD tools available that would be able to do this,
obviously all the major CFD engines would do this comfortably (Airpak from
Ansys Fluent most easily I suspect with its integrated solar gain modelling
tool) but they are going to be quite expensive and there is quite a learning
curve to some of them.

Alternatively, you could use something like PyroSim, which is a good front
end for NIST that would probably be suitable. NIST does transient analysis
and is used commonly for smoke/fire analysis. It uses LES turbulence models
from memory and so would be likely to be more accurate for the sort of
stratified flows you?d get in an Atria as well. You could then set
transient boundaries based on files of data exported by the dynamic engine.
Might take a bit of work, but it should work.

Regards

Dr Paul Carey

Offline
Joined: 2011-10-02
Reputation: 0

Frederik, I have worked on a couple of large atrium projects over the years. You did ask for comments and suggestions so here goes:

You need a tool that can predict conditions in a quite complicated space and looking at you location you will have two major problems:

1. downdraft and cold radiation during the winter
2. high radiation surface and possible stratification during the summer.

I model the space with ROOM (ARUP Rand D, Mike Holmes and Andrew White) or lately TRNSYS. This enables me to set up heat balances and run fairly quickly (once the models have been built), and provide what if scenarios for both summer and winter. ROOM enables me to look at radiation exchange in the space and occupant comfort, as well as stratification. TRNSYS will allow analysis of air movement within the space, this is especially important when heating the space in winter.

Once these have been completed you can step to CFD (Colours for Directors), at this point I will probably get shot down, but most of the CFD programs I know are steady state, so for one moment in time. The next problem is the complexity of the space, the grids and the number of cells all of which increase the validity of the results and drastically increase computational time. For example CFD runs for the atrium in the LA courthouse were between 56 to 72 hours each.

I think you need to work out what results you require and the level of accuracy of these results. Is this a design problem or an academic project?

We published a paper on temperatures and bulk air movement for a large atrium in a hospital in the Netherlands, which had a large glazed roof and numerous openings, this was about 15 years ago and is published in ASHRAE transactions. I also presented a paper on a large atrium/shopping mall in Warsaw at the opening of DTU new labs in 2001. The LA courthouse data has never been published as it is the property of the client, but I may be able to find some of the results for you. To give you some idea of the costs and time involved for this particular analysis, the CFD analysis alone was in the order of \$120,000 and our effort, temperature, air movement, thermal comfort and stratification as well as writing a report, that could be understood by the client another \$240,000 and we didn't make any profit.

I hope this helps, if not let me know what else you need, good luck.

Peter Simmonds. Ph.D.

Offline
Joined: 2011-10-02
Reputation: 0