First time here? Check our help page!
3

E+ Unmet hours when delta temperature is below reporting tolerance?

I've got a VAV system with Reheat serving multiple zones: 4 offices and a corridor on the same floor. The corridor is producing unmet hours during nighttime operation.

  • My Tolerance for Time Heaitng Setpoint Not Met is 1K (1.8R).
  • The setpoint is 14.44°C // 58°F.
  • While the zone temperature is around 14°C, so that $\Delta T < ReportingTolerance$ I still get unmet hours.

See this graph, where:

  • Black is Zone Air Temperature
  • Light blue/turquoise is Zone Thermostat Heating Setpoint Temperature
  • Darker blue/marine is Zone Heating Setpoint Not Met

Unmet Hours while below reporting tolerance

What's going on?

Bit of context: this zone is served together with 4 others by a VAV system for which there's an AvailabilityManager:NightCycle (thermostat tolerance is 1°C) and Outdoor Air is turned off between D at 10pm and D+1 at 6am. Two zones out of 5 do produce more unmet hours than the others (one office and this one being a corridor).

Julien Marrec's avatar
29.7k
Julien Marrec
asked 2016-02-18 03:16:04 -0500
edit flag offensive 0 remove flag close merge delete

Comments

add a comment see more comments

1 Answer

4

I think you are seeing the impact of the 1 C tolerance in the AvailabilityManager:NightCycle. Try a zero tolerance there.

Archmage's avatar
7.2k
Archmage
answered 2016-02-18 09:35:50 -0500
edit flag offensive 0 remove flag delete link

Comments

Per the docs, 1°C tolerance would just mean the system would kick on at 14.44-1/2 = 13.94C. Am I missing something?

Julien Marrec's avatar Julien Marrec (2016-02-18 10:06:21 -0500) edit

Is your Zone Air Temperature reported at the detailed frequency? You want some cushion between the night cycle offset and time-not -met reporting tolerance, they are both 0.5K.

Archmage's avatar Archmage (2016-02-18 15:44:25 -0500) edit

That's a very good point, as usual. It was indeed reported at the Timestep frequency, and there could indeed be stuff hapening that I wouldn't see without going to the detailed one. But my reporting tolerance is 1K, and the night cycle offset is 0.5K, technically I still have 0.5K cushion?

I'm trying to reproduce my "problem" on a small model, because my actual model takes an hour to run and adding a bunch of variables at the detailed timestep won't help with runtime...

Julien Marrec's avatar Julien Marrec (2016-02-19 03:25:37 -0500) edit
add a comment see more comments