Hi Peter,

Thanks for your reply on the matter. I have began looking a little at TRNSYS, and from what you are describing about its air movement analysis, it sounds tempting to use this program for the dynamic simulation. How does TRNSYS model the temperature stratification in the atrium? Can it predict the air flow if the solar gain is "captured" by interior screens at the top of the atrium, (which thereby adds to the air movement at this location) instead of the solar gain being added near the walls and floors by tracking the solar irradiation? What about diffuse radiation? In short: What is your experience using TRNSYS? Can it do what you want it to do? What about stratification?

My client expects presentation of results regarding temperature in the occupied zone as well as air velocity. However I will not make a CFD stating that the air velocity on the occupied zone is 0,1435 m/s (...which is shown in this CFD calculation etc. ...) because the calculation is only steady state and is not valid for a whole year, because of the uncertainties in the input in the CFD calculations. Moreover I can't present the results from the thermal calculations as these are very uncertain due to the uncertainty about stratification in the atria. Can TRNSYS take care of these problems?

I would rather not make more than 4 CFD simulations, and the simulations should have the necessary input from TRNSYS in order to achieve the correct results more or less in the first simulation. What's your comment on transient simulation? Is it worthwhile? From what I know of these simulations, the computing time is more or less the same as the results are based on the constant boundary conditions and is therefore only showing the uncertainties in the solving of the different differential equations. The results presented in the transient calculations can then be used to illustrate the variation of the results... or can it?

The CFD calculations are extremely expensive and if it is to be of any use for any client the uncertainties should also be presented. Your efforts on LA courthouse confirms that CFD is expensive. When was this work done? What was the tasks on this building? It would be nice if I could glance at the papers you made at that time.

I haven't been able to find the program "ROOM"? Do you have a link? Thanks for your help.

Kind regards

Frederik Vildbrad Winther (FRW)

Hi Paul,

Thanks for your reply. I haven't heard about DesignBuilder CFD before and I have only heard and peaked to EnergyPlus. EnergyPlus seems OK for simulating multizones and thermal simulations. Do you know how it handles stratification? I am thinking of using thermal simulations, with a good link to a multizone program so that the 2 programs can predict the air movement as well as the stratification in the room. I am then thinking of doing 4 CFD studies and these simulations should have input from the thermal simulation. The output from the CFD simulations should then be compared to the results from the thermal simulation. The two calculations should therefor be reasonably comparable (in the end).

I think that the use LES turbulence model for indoor climate analysis is rather hazardous because this turbulence model is mostly used for fire and smoke simulations. I have obtained good results for indoor climate using ke-turbulence model. Is my thought "out of date" on this point? Does ke-turbulence model also work for larger rooms? When can/should LES-turbulence model be used, apart from smoke/fire simulations? The problem is also getting a thermal simulation program, which can import the results from a CFD calculation.

Using the transient calculations could prove very useful in showing the uncertainty in the CFD calculations. However, can I, when importing the boundary conditions from the thermal simulation to the CFD simulation, import the temperature stratefication as a start condition for the simulation? What is your experience on coupling thermal/multizone simulations with CFD calculations? Does it work? It could be very useful if the CFD calculation starts with the same boundary conditions and results from the thermal simulation, rather than starting from scratch every time a CFD analysis has to be performed.

Do you know of any papers dealing with modeling of solar gain in CFD tools?

Med venlig hilsen

Frederik V. Winther

Frederik,

I have done bigger building than this. You need to run two worst?cases,?cooling and heating. You will see the stratification in cooling case and possible down draft (or no) in heating. The critical part of your problem is the radiation. You need to include this effect for sure. As for the turbulence model, I recommend you to use indoor zero equation.

Thanks

Ding Li, PhD, PE

I put my comments in blue below.

Ding Li, PhD, PE

Hi Ding,

With regards to importing the boundary conditions from the dynamic (thermal) simulation, how do you take building mass into account. Some of the external conditions i absorbed by the floors and walls. Can you get the CFD calculations to take this factor into account when using CFD? I haven't heard of this before.

Also the CFD results are only a predicted airflow and are extremely sensitive to input parameters, and the flow is more or less steady state. Therefore the CFD can only be used to illustrate what could possibly happen, and to get an idea of the stratefication. However as you mention with regards to stratefication, I need a dynamic simulation which can calculate stratefication relatively simple so that the stratefication would vary over time. E+ sounds like a possibility, but if it only can calculate stratefication when using displacement ventilation, then it becomes less attractive to use. Do you know of any other software which can calculate stratefication? The Danish program BSim has a fixed parameter for stratification so this program can not be used for calculation over a whole year unless the calculation is divided up into smaller parts, and the stratefication is changed manually, based on CFD calculations.

I agree with you that it is important to include radiation! However it takes longer to obtain convergance with this module "on". How do you manage to calculate large rooms with many surfaces, and still including radiation? Clustering of (many) computers?

Med venlig hilsen

Frederik V. Winther

Frederik,

You are thinking too much. IMO, you don't need to do dynamic simulation. You only need the worst case. If you want to do CFD simulation for a year, it's almost impossible. It takes forever for the calculation to finish. The industry don't even calculate a year load to size the equipement, why do you need to? Unless you want to do energy analysis for the building for the whole year, you are not supposed to do that.

If you want to do energy modeling for the building for a year and you want to include stratification effect, I know E+ and Trace have DV feature. Actually Trace just released this feature about a month ago since it produced unexpected result. You may try these two products.

As for the stratification, it's not?very sensitive to your input. If you create the model correctly and the computation converges, the stratification is quite stable.

About the radiation, you need to choose the appropriate model. The calculation time can be much longer if you have many objects in the model. It's easy to take several days for one CFD model to finish?from my experience.

Ding Li, PhD, P.E.

Hi Ding,

I need to do a dynamic simulation, because the client wants to know how many hours during the year, his demands regarding indoor climate are met. However I won't do CFD calculations for a whole year, because of the computing time. It's simply not possible. This is why I need a dynamic simulation tool which can simulate varying temperature stratefication and different loads. Your advice on using E+ is therefore taken ito concideration as well as TRNSYS.

Both programs offer som output which can be imported into the CFD calculations in order to glance at the draft rate in the occupied zone. For the CFD calculations I will use your advice on using "Indoor zero equation" and see how it works, as well as minimizing the amount of objects in the model so that I can use the radiation module.

Thanks for your help and advice. Hope it works. :-)

Kind regards

Frederik Vildbrad Winther (FRW)

Good luck on your work.

Ding