What output variable will show a difference between a model with convex vs. non-convex surfaces

asked 2018-09-04 18:41:10 -0600

mldichter
2575 ●2 ●11

updated 2018-09-05 14:09:36 -0600

I have a model with some non-convex surfaces and I receive this warning

   ** Warning ** DetermineShadowingCombinations: There are 12 surfaces which are receiving surfaces and are non-convex.
   **   ~~~   ** ...Shadowing values may be inaccurate. Check .shd report file for more surface shading details
   **   ~~~   ** ...Add Output:Diagnostics,DisplayExtraWarnings; to see individual warnings for each surface.

I have changed the non-convex surfaces to convex surfaces in the model. What output from Energyplus will show a difference in the simulation, if any, between the model with non-convex surfaces and the model with all surfaces converted to convex shapes?

I want to use the model with the non-convex surfaces since it runs faster, but not at the expense of incorrect results, so I want to compare the output of the two simulations.

UPDATE:
I split up 2 of the 12 non-convex surfaces into multiple convex surfaces. One surface I split was interior to the building and it added about 0.02 GJ to the energy usage. The other surface I split was a flat roof which, strangely, decreased the energy used by about 0.7 GJ. I thought changing the roof would have no effect because it is a horizontal roof that is not being shaded.

image description

@Julien Marrec I am unaware of which solar distribution I am using. The field is blank. What is the default?

image description

answered 2018-09-05 05:07:34 -0600

Julien Marrec

29716 ●94 ●76 http://www.effibem.com/

updated 2018-09-06 02:32:52 -0600

I don't think there's a need to go crazy here. Just start with total annual site energy consumption. If that's off, you can stop there, you should use the non convex one.

If it's very close, just make sure you aren't trading heating for cooling or vice versa by comparing monthly total site energy consumption.

Edit: to answer your update post.

Finding Default values:

Most of the time you can find the default for a field by looking at the I/O reference guide. In the case of Solar Distribution, the I/O guide here doesn't give the default. Then you're left looking a the Energy+.idd (it's in your E+ installation folder), which gives you this:

  A3 , \field Solar Distribution
       \note  MinimalShadowing | FullExterior | FullInteriorAndExterior | FullExteriorWithReflections | FullInteriorAndExteriorWithReflections
       \type choice
       \key MinimalShadowing
       \key FullExterior
       \key FullInteriorAndExterior
       \key FullExteriorWithReflections
       \key FullInteriorAndExteriorWithReflections
       \default FullExterior

So the default is FullExterior, which means there is self shading and all beam solar radiation is assumed to fall on the floor. It's probably an adequate choice to "minimize the accuracy of non-convex surfaces" like you said.

Results of your try:

You're seeing a -0.0026% difference in EUI, which is basically peanuts, especially compared to the actual uncertainties in a model. I'd say it's probably fine to use the non-convex model.

edit flag offensive delete link

Comments

Just a hunch: for 12 non-convex surfaces, I doubt you'll see much difference unless they are huge. Also, the effect will depend on the solar distribution you are using.

Please report back (just a short comment will do) once you check it, I'm curious.

Julien Marrec ( 2018-09-05 05:09:01 -0600 )edit

@Julien Marrec I tried out changing a couple of the surfaces to convex. (see edited post)

As for the choice of solar distribution, the field is blank so whatever the default is. What choice of solar distribution would minimize the inaccuracy of non-convex surfaces?