Fix #11039: supp heat blocked by negative QToHeatSetPt#11546
Fix #11039: supp heat blocked by negative QToHeatSetPt#11546brianlball wants to merge 7 commits intodevelopfrom
Conversation
Unit test proves supplemental heating is blocked when QToHeatSetPt < 0 despite zone demand exceeding primary coil capacity. The check at controlUnitarySystemtoLoad (~line 8125) compares FullSensibleOutput against QToHeatSetPt — when the zone predictor underestimates heating need, QToHeatSetPt goes negative and backup never fires. Test will pass once the fix is applied.
…into issue-11039-supp-heat-blocked
Supp check used QToHeatSetPt as threshold — when negative (predictor overestimates zone temp), positive coil output always passes check, blocking supp. Two fixes: 1. Constant-fan ZoneLoad overwrite (lines 10808/10830/10856/10878): use max(QToHeatSetPt, QZnReq) to preserve original demand signal. Previously missed the DualHeatCool deadband case at line 10878. 2. Supp check zero-floor: when HeatingLoad is true but suppLoadTarget is negative, floor at 0 so supp fires when primary can't achieve neutral output. Handles SingleHeat where QZnReq==QToHeatSetPt. VSHeatPump integration test: 53.25 → 15.00 unmet heating hours (-72%).
New test: SuppHeatBlocked_ConstantFan_DualHeatCool — continuous fan, DualHeatCool deadband, QToHeatSetPt=-500, QToCoolSetPt=3000. Verifies supp fires via the zero-floor fix when primary targets negative load. Fails on unfixed code, passes with fix. Reverted line 10878 (DualHeatCool deadband) from max(QToHeatSetPt,QZnReq) back to QToHeatSetPt. The max() produced ZoneLoad=0 which triggered the small-load guard at line 10942, killing HeatingLoad. The supp check zero-floor handles this case instead.
…into issue-11039-supp-heat-blocked
brianlball
added
Defect
|
|
|
Is there a defect file to test? I won't believe this change unless I see a defect file's result change for the better. |
mitchute
added
DoNotPublish
|
|
max(QToHeatSetPt, QZnReq) produced ZoneLoad=0 in deadband, triggering the small-load guard that cleared HeatingLoad. Supp heat zero-floor at line 8207 is sufficient alone.
|
|
|
|
|
At 14:00 on Jan 1, why is speed 2 out of 4 on and the zone temperature is below the heating set point? The supp coil is on here but there is more capacity available from the main heating coil. It seems like there is something wrong with the multispeed coil and onlhy possibly the logic for the supp coil. 
This branch shows the heating coil and supp heater are on when they should not be. The zone temp is forced to the cooling set point when the heating coils should not be on when the zone temp is above the heating set point temperature. 
|
|
@rraustad Thanks for your comment in the issue. I had just assumed the coil was maxed out and why the supp heat was needed. Thats clearly not the case and my issue diagnosis was wrong. Something else is going on. I'm still new to this part of the code base and trying to figure things out, but I think |
|
Here's a figure I've been working on to help explain the difference between QToHeat/CoolSetPt and ZoneLoad. ZoneLoad starts out as the raw QToHeat/CoolSetPt load (comes from RemainingOutputRequired or SequencedOutputRequired from predictor) but can switch signs if there is continuous fan operation and depends on the load imposed by outdoor air. If there is a HeatingLoad then ZoneLoad can be positive or negative, depending on the outdoor air conditions when using constant fan. The red text is what happens when OA causes the load direction to change. If there is a negative ZoneLoad when HeatingLoad = true, that usually means the zone temp will wind up at the heating SP temperature (similar logic for cooling). In this case the system is actually providing cooling to the zone but the heating coil (or supp heater) is on, not the cooling coil. The system is just not providing excess cooling to push the zone temperature below the heating set point temperature. The 2 Q loads shown in (xxx) at the right of the figure just show that it's possible to have either sign for these Q variables based on what predictor calculates. Black is no system operating and red is system operating with no coils active and only the fan is operating. Coils will be activated later after ZoneLoad is determined and HeatLoad/CoolLoad are set. This happens in initLoadBasedControl inside this conditional: 
|
4 participants
Summary
QToHeatSetPtnegative. The supp check (FullSensibleOutput < QToHeatSetPt) always passes when QToHeatSetPt is negative.initLoadBasedControloverwritesZoneLoad = QToHeatSetPt, making the primary coil target the wrong (negative) load. The supp check then sees the primary "met" the target and doesn't fire.ZoneLoadatmax(QToHeatSetPt, QZnReq)in 3 constant-fan heating paths to preserve original demand signal; (2) floorsuppLoadTargetat 0 whenHeatingLoadis true, so supp fires when primary can't achieve neutral output.Test plan
SuppHeatBlocked_NegativeQToHeatSetPt— cycling fan, SingleHeat, negative QToHeatSetPt. Fails on develop, passes with fix.SuppHeatBlocked_ConstantFan_DualHeatCool— constant fan, DualHeatCool deadband, negative QToHeatSetPt. Fails on develop, passes with fix.