E+ Space FT Issue

[shorowit]

Issue overview

I'm doing some testing of the new E+ Space FT on residential models. Since the models have 1 space per zone, I expect identical results between OS 3.4.0 and 3.4.1.

Current Behavior

For a sample file, I'm getting different E+ results due to the ZoneInfiltration:DesignFlowRate object changing.

Expected Behavior

The ZoneInfiltration:DesignFlowRate object should be identical between OS 3.4.0 and 3.4.1

Steps to Reproduce

Using OpenStudio-HPXML, run:
openstudio workflow/run_simulation.rb -x workflow/sample_files/base.xml
in both OS 3.4.0 and 3.4.1 and note the different E+ results.

(To run OS 3.4.1, you need to change this line to "3.4.1".)

This is caused by a different ZoneInfiltration:DesignFlowRate object:

   Always On Discrete,                     !- Schedule Name
-  AirChanges/Hour,                        !- Design Flow Rate Calculation Method
-  ,                                       !- Design Flow Rate {m3/s}
+  Flow/Zone,                              !- Design Flow Rate Calculation Method
+  0.00922493533933472,                    !- Design Flow Rate {m3/s}
   ,                                       !- Flow per Zone Floor Area {m3/s-m2}
   ,                                       !- Flow per Exterior Surface Area {m3/s-m2}
-  0.1,                                    !- Air Changes per Hour {1/hr}
+  ,                                       !- Air Changes per Hour {1/hr}
   1,                                      !- Constant Term Coefficient

If I manually change the IDFs so that this object is identical, then the E+ results match perfectly.

[joseph-robertson]

Looks to be due to this: https://github.com/NREL/OpenStudio/blob/develop/src/energyplus/ForwardTranslator.cpp#L378-L385

[jmarrec]

My hunch is that the geometry in question is using a space that has non rectangular geometry and since SpaceInfiltrationDesignFlowRate::hardSize() when going from ACH uses the calculated Space::volume:

https://github.com/NREL/OpenStudio/blob/d39d697a2aa97984aeafeada0ec7563435632658/src/model/SpaceInfiltrationDesignFlowRate.cpp#L123-L126

I think it's #127 that's causing it

[jmarrec]

I'm a bit confused by the geometry I find in OpenStudio-HPXML/workflow/sample_files/run after running openstudio workflow/run_simulation.rb -x workflow/sample_files/base.xml

[shorowit]

HPXML files don't include full 3D building geometry (just surface areas and orientations), so we "explode" the surfaces in order to prevent self shading and so you can better see the surfaces (relative to a pinwheel result where all of the surfaces go through the origin). If I recall correctly, this is is the same approach used by CBECC-Com (using OpenStudio/EnergyPlus) for one of their simplified geometry options. EnergyPlus certainly supports non-enclosed building geometry.

[jmarrec]

I guess the technically required cases where we need to hardsize the SpaceInfiltration are:

      for (auto& infil : model.getConcreteModelObjects<SpaceInfiltrationDesignFlowRate>()) {
-       // TODO: technically we only need to do that if the space it's assigned to is part of a thermalzone with more than one space
-        // Same reason as above: not doing it for now
+       // technically we only need to do that if the space it's assigned to is part of a thermalzone with more than one space
-       infil.hardSize();
+       if (infil.space() && !openstudio::istringEqual("Flow/Area", infil.designFlowRateCalculationMethod()) && infil.space()->thermalZone().spaces().size() > 1) {
+         infil.hardSize();
+       }
      }

https://github.com/NREL/OpenStudio/blob/d39d697a2aa97984aeafeada0ec7563435632658/src/energyplus/ForwardTranslator.cpp#L378-L385

[jmarrec]

Side note, but #4579 is also a good workaround this issue for your model. If you hardcode space.setVolume(xxx) then the conversion ACH => Flow/Space will work too

[joseph-robertson]

I guess the technically required cases where we need to hardsize the SpaceInfiltration are:

      for (auto& infil : model.getConcreteModelObjects<SpaceInfiltrationDesignFlowRate>()) {
-       // TODO: technically we only need to do that if the space it's assigned to is part of a thermalzone with more than one space
-        // Same reason as above: not doing it for now
+       // technically we only need to do that if the space it's assigned to is part of a thermalzone with more than one space
-       infil.hardSize();
+       if (infil.space() && !openstudio::istringEqual("Flow/Area", infil.designFlowRateCalculationMethod()) && infil.space()->thermalZone().spaces().size() > 1) {
+         infil.hardSize();
+       }
      }

https://github.com/NREL/OpenStudio/blob/d39d697a2aa97984aeafeada0ec7563435632658/src/energyplus/ForwardTranslator.cpp#L378-L385

Note I think you meant "Flow/Space" (instead of "Flow/Area") above.

[jmarrec]

absolutely, Flow/Space