Can you clarify what EnergyPlus version you are using to simulate the .idfs on the DOE website?

One consideration is that the default files do not include an OutputControl:ReportingTolerances object, which specifies the temperature tolerance reporting for unmet hours. The default value is 0.2 deltaC, 0.36 deltaF, which a very tight tolerance.

Simulating the 90.1-2010 Large Office in Buffalo in EnergyPlus version 8.2.0 I get these unmet hours: default 0.2 deltaC / 0.36 deltaF tolerance: 286 occupied unmet hours heating, 677.5 occupied unmet hours cooling more appropriate 0.555 deltaC / 1 deltaF tolerance: 185.25 occupied unmet hours heating, 286.25 occupied unmet hours cooling

Another consideration is that EnergyPlus reports occupied unmet hours if there is any occupancy in the zone, and many schedules shut the HVAC system off below 5% or 10% occupancy. So these get reported as occupied unmet hours even though the HVAC system is intended to be off.

In general, unmet hours <300 hours is acceptable - though it's always good to check how far of the zone is from setpoint.

For more information on the DOE prototype buildings, see the answer to this question about the DOE Prototype buildings.

The posted DOE Prototype Models in EnergyPlus are in Version 8.0. When we complete the ASHRAE 90.1-2019 analysis, the DOE Prototype models will be reposted in Version 9.0 (or later). The 2018 IECC models are under review and will be posted when the review is complete. Matt is right that the occupancy schedule BLDG_OCC_SCH includes some not-zero fraction in hours when the fan is scheduled off (HVACOperationSchd ). Those non-zero occupancy hours are meant to represent housekeeping crew activities. When the night setback setpoint is not met by AvailabilityManager:NightCycle, EnergyPlus reports those hours as Time Setpoint Not Met During Occupied Heating/Cooling.
A couple of more reasons that could contribute to the Not Met hours during the system startup and daytime hours 1. Sizing factor The sizing factors for most of the DOE Prototype models (except for Secondary School, which probably should be fixed) with multiple zone VAV systems are purposely set to 1.0 for both heating and cooling. Ideally, they should follow 90.1 App G rule, i.e. oversize by 15% for cooling and 25% for heating. However, the rule is intended to oversize heating and cooling coil capacity only, not for airflow. EnergyPlus currently doesn’t support that and would oversize both coils and airflow, which causes the hourly fan airflow rarely reach its design size and also the minimum VAV box flow (say 30% of oversized box size) higher than necessary. 2. Setpoint manager EnergyPlus V8.0 cannot always maintain the setpoint of multiple zone VAV systems when SetpointManager:OutdoorAirReset or SetpointManager:Scheduled is used. When supply air temperature is lower, the reheat coil may not have enough capacity to meet the heating load. I believe the setpoint managers perform better in newer versions of EnergyPlus.

Regarding the observation that 90.1-2013 Large Office Prototype uses more energy than 90.1-2010, that’s because Addendum BT to 90.1-2010 (PDF page 206 https://www.ashrae.org/File%20Library... ) expands the ERV requirements and triggered the Large Office prototype in CZ 4A to have ERV. Ideally, that should save energy but Section 6.5.3.3 Multiple-Zone VAV System Ventilation Optimization Control in 90.1-2013 exempts AHU with ERV to have dynamic ventilation control. So, the 90.1-2013 Large Office model with added ERV and removed ventilation optimization control uses more energy than 90.1-2010. The 90.1 committee fixed this problem in 90.1-2016 by removing that exception, see Addendum J to 90.1-2013 (PDF page 12 https://www.ashrae.org/File%20Library... or PDF page 54 of the Final Energy Savings Analysis ANSI/ASHRAE/IES Standard 90.1-2016 report https://www.energycodes.gov/sites/def...). The EUI of 90.1-2016 Large Office is the lowest among all versions of 90.1 for all climates. Thanks