what is the maximum file size?

[anthony75]

Hi
I am attempting to test thermal loading reduction due to tree shade and need to compare the difference between a detailed tree with a simple sphere. I know it's probably rediculous but my file size is 54MB. Just opening the file takes about 30 minutes. I expect to leave it running for days or weeks so thats ok but I still need to do a bit more modelling and it's just not coping. Is this file size to big or is my computer underperforming? I have a quad core with 4GB ram and 200GB free space. I am also running the file from the c drive.
Thanks
Anthony

[JosephG]

The file is too big.

[PCully]

Yeah I agree 54MB sounds like you have a lot of surfaces in there. I'm not surprised the <VE> is starting to slow down at this size of project and can't guarantee Suncast will even be able to allocate enough memory to run by the sounds of it. Feel free to try though.

If you are finished working on the Shading surfaces then you could try creating a project with only the shading geometry then export this as a GEM file and import to the original project when it is fully built and you're ready to start the analysis. This would speed things up but as I say I can't guarantee how the analyses will run when the project is complete.

Good luck,
Phil

[anthony75]

Thanks for the replies.
Can you suggest an alternative means of simulating the difference between varied tree canopy density? I thought there would be a way to assign transmittance values to 2D shading planes but I have not had much success here.
I modeled a 2D surface with holes cut out to simulate a shading coefficient but this seemed to produce a file of similar size to the 3D trees.
Any suggestions would be appreciated.
Thanks again
Anthony

[PCully]

I'm not aware of any study like this having been performed in the <VE>, we would normally advise users to approximate the shading provided by surrounding foliage so they can avoid the issues you are now seeing when you have too many shading surfaces.

If you do persevere with this I would be interested in the kind of variation you see between different densities of shading surfaces compared to using a single shape.

Phil

[anthony75]

Hi Phil
I have had some success eventually.
I ended up modeling the trees as 2 circular planes intersecting at right angles. For the transparency I ended up putting holes in the planes so you would still get a similar diffuse lighting. As a test I modeled identical size trees without the cutouts as well as a building with no shading at all.
Suncast eventually processed it after about 3 days of running (a single pass).
Apache runs a simulation in 20 seconds or so now that the shading calculations are out of the way and it looks good. Results are not great but it's clear that the different shading factors is reflected. I'm getting a reduction in cooling loads of 3.3% and 5.3% for the two tree types when compared to the treeless control.
I have been trying to get the Surface temperature to show results in Vista Pro (Beta) but I assume that feature is still to come. It would be great as we have thermocouples and heat flux sensors monitoring our buildings constantly so comparisons would be helpful.
There have been a few studies in California on this with simulations run in DOE-2.1. I don't know anything about that software but their results were underestimated by about 50% when compared to physical studies.
To give you a bit of background, I am an master of architecture student at Melbourne University, Australia. I'm studying the effect of tree shade on thermal loading of buildings and the potential for cooling energy reductions. I'm also considering evapotranspiration but I am not attempting to model that.
We have are running a controlled full size physical experiment along side the simulations for at least a year.
So my issue has been to model the shading effect of native Eucalyptus vs a London plane tree.
I will post a link to our blog if results start lining up with measured data.
Sorry for the long reply!
Anthony

[ZapBran]

enjoying this post guys - we might be in danger of a community breaking out.
It would be good if we could specify a shading device and assign a profilethat determined the opacity... 1=fully opaque and 0=fully transparent then we could handle deciduous trees as well as... oh... glazed canopies etc.

[EcoNerd]

Yeah it would be great to be able to simulate buildings surrounded by deciduous trees. Is there nothing like it yet? I think it would even be quite important to have a function for it considering the positive effect of decicuous trees in moderate climate...

[huntervu28]

@ZapBran : strictly, we can do it

My Idea: Modeling a room in a form of flat tree (thinner is better) then drawn 100% window to the 2 main big face. Assign clear glass for one windows, the 2nd should be assigned glass with blinds and profile for winter summer (in material editor, evidently). Don't forget exclude all "tree room" out of any energy systems. "Tree room" type can not be changed to any shading room type, IES will not treat blinds inside glass.

ps: I have no time to test, may someone do it and post results.

[antonin-peter]

Hi everybody,
Since January 27th, is there any other way to model a deciduous tree? For instance, by associating a schedule to local shade (which doesn't seem to be obvious)?
Thanks by advance,
Antonin