Design conditions incorrectly parsed for TMYx EPWs

[shorowit]

Issue overview

TMYx EPWs have different numbers of values for DESIGN CONDITIONS in the header relative to TMY3 EPWs. If you compare, e.g., this TMYx and this TMY3, you'll find:

• Heating: 15 values (TMY3) vs 16 values (TMYx)
• Cooling: 32 values (TMY3) vs 32 values (TMYx)
• Extremes: 16 values (TMY3) vs 15 values (TMYx)

It appears that OpenStudio reads the values by position, so this causes all the cooling design values to be assigned to the wrong property. For example, the 0.4% MCWB is assigned to the 1% Cooling DB.

Instead of reading the data positionally, OpenStudio should read the data relative to each "header" (i.e., Heating, Cooling). This would be more robust.

This issue causes OpenStudio-HPXML to use 74.4F as the cooling design temperature instead of 90.7F for the above TMYx, resulting in very wrong HVAC autosizing results for cooling.

[shorowit]

The list of possible values appears to be here: https://energyplus.readthedocs.io/en/stable/auxiliary-programs/auxiliary-programs.html#design-conditions-header-data-csv

[joseph-robertson]

Ideas:

Option 1

1. Count number of positions after Heating, Cooling, Extremes
2. If Heating has 16, then we know WSF is the last position
3. If Extremes has 16, then we know MaxWetbulb is in position x
4. If Heating/Extremes neither 15 or 16 (or Cooling not 32), don't return any design conditions?

Option 2

Use the "Title of Design Condition" of the design condition source (e.g., if has "ASHRAE", then build heuristics based on year, like 2009) to set the expectations.
We may need to do a search to understand better all the design condition sources.

[jmarrec]

I have reported and fixed similar issues for parsing stat files, which is done by openstudio-standards:

• Stat file throws when parsing weather with no dry period openstudio-standards#1946
• And more importantly this, which is specific to TMYx and is the same thing: Cannot parse design_temp_info for stat files from climate.onebuilding.org openstudio-standards#1948