Unmet hours too high in annual simulation but 0 on design days EnergyPlus

edit

It is difficult to pinpoint an answer given this much information, but here are some possible reasons to explore:

1. The summer design day is of the type that has high drybulb with mean coincidenct wetbulb, but often cooling coil loads are larger for high wetbulb with mean coincident drybulb. The program will select the design day with the highest sensible zone load, so sometimes you have to remove the DB->MWB day from input to force it to use the WB->MDB day. This often comes up for dry climates, which I would think is the case for Lima.

2. The schedules that are important for controls are setup one way for the design days, but then differ for the regular types of days so that the system is not really running properly when changing from design days to annual run.

3. Sun angles for a mid-summer are high, but shoulder seasons have lower sun angles which actually lead to higher gains in the space, regardless of outdoor temperatures. Add some additional design days that cover shoulder seasons, even covering all twelve months.

As always, you'll want to study detailed timeseries results for air system nodes to diagnose what is really going on during those hours not met.

3000 unmet hours is clearly too many, because that 's almost 35% of the entire year, so obviously something is seriously wrong here. I can't tell from your description the difference between the two models - are they the same building, only one with autosized equipment and the other with capacities specified by the engineer, or is one a reference building w/autosized equipment and the other the candidate building with predefined equipment capacities ? If the former, then you can just compare the autosized capacities to the predefined ones and see how much is the undersizing. If the latter, then you can't compare the capacities (because the building loads are different), but you can bump up the autosized capacities and watch how the number of unmet hours decrease. I'm not certain, but I believe that EnergyPlus allows input for a Sizing Factor to make it easy to do this.

However, before doing any such exploration, make sure that your design calculations are not faulty - (1) Is the DDY file for the same location and how do the design conditions compare to the maxima on the weather file, (2) is the right month being used to calculate the clear sky radiation? (it should be either January or February for Lima, not July or August as typically used for North America).