Chiller on SetpointManager:FollowSystemNodeTemperature reports error with no cooling load

asked 2018-05-21 21:21:41 -0600

updated 2018-05-22 07:37:44 -0600

I am experimenting with using a setback scheme for supply air and chiller water temperatures, using Energy Plus's standard "RefBldgLargeOfficeNew2004_Chicago_SA_setback.idf" building model with weather file "USA_IL_Chicago-OHare.Intl.AP.725300_TMY3.epw". This is in Energy Plus 8.3.

The plan is that the supply air temperature setpoint will be regulated to match the warmest zone's requirements, rather than always set at 12.8 deg C. Then the chiller water temperature setpoint will be follow at 6.1 deg C below the supply air temperature setpoint, rather than being held at 6.7 deg C all the time.

The idea of this strategy is that the chiller water can rise to a higher temperature when there is low cooling load, increasing the efficiency of the chiller. In addition, if the zone air setpoints are raised, the setpoints for the supply air and chiller water will follow, and the chiller will reduce its work or even shutdown for a while. This would be useful for cutting electricity use and "coasting" in response to a short period of high electricity prices.

Concretely, I'm planning to implement this strategy in Energy Plus by replacing the SetpointManager:Scheduled controllers with"SetpointManager:WarmestTemperatureFlow for the VAV outlet nodes and SetpointManager:FollowSystemNodeTemperature for the chillers. The chillers' setpoint managers will be indexed to the temperature at the VAV_1 Supply Equipment Outlet Node, since there doesn't seem to be a way to index them to the VAV_* SAT setpoints themselves.

As a first stage, I have inserted the SetpointManager:FollowSystemNodeTemperature controllers in place of the SetpointManager:Scheduled controllers for the chillers. However, now when I run Energy Plus, I get warning messages like this:

 ** Warning ** CHILLER:ELECTRIC:REFORMULATEDEIR "COOLSYS1 CHILLER 1": The part-load ratio (0.000) is outside the range of part-load ratios (Y var) given in Electric Input to Cooling Output Ratio Function of Part-load Ratio bicubic curve = REFORMEIRCHILLER CARRIER 19XR 1259KW/6.26COP/VANES EIRFPLR
 **   ~~~   ** The range specified = 0.180 to 1.030. Environment=CHICAGO ANN CLG .4% CONDNS WB=>MDB HVAC Sizing Pass 1, at Simulation time=07/21 00:10 - 00:20

I'm not surprised that the part-load ratio would drop to 0.000 under this control strategy, but I am concerned that Energy Plus isn't recognizing this and just turning off the chillers. I think it is handling part-load ratios between 0 and 0.1 by cycling the chillers and between 0.1 and 0.2 by hot gas bypass (using the 0.2 point on the performance curve). In the default setup, I think it also manages to turn the chillers off on cool days without complaining. But for some reason, when the SetpointManager:FollowSystemNodeTemperature calls for no chilling, Energy Plus reports these errors, rather than recognizing this special case and shutting off.

Is there anything I can do to avoid this message? Or should I just ignore it? Or follow some different strategy entirely?

edit retag flag offensive close merge delete

Comments

Could you share your Chiller:Electric:ReformulatedEIR input? If you didn't modify the example one, I thinkt he Minimum Part Load Ratio might be 0.1, while your curve Minimum Value of y appears to be 0.18. (Granted that still wouldn't explain why you get the warning at a PLR of 0.00 though)

Julien Marrec ( 2018-05-22 03:47:23 -0600 )edit

@Julien-Marrec Yes, the only change from the standard example is the SetpointManager:FollowSystemNodeTemperature that I added. So yes, the minimum part load ratio is 0.1, but the Energy Plus documentation seems to say that the chiller will run on hot-gas bypass between 0.1 and 0.2 load. I presume that uses the energy requirement for the 0.2 level. At any rate, I'm wondering why it complains about running at 0 load, instead of just shutting down (or 'cycling' at 0%)?

Matthias Fripp ( 2018-05-22 19:46:40 -0600 )edit

add a comment

2 Answers

Sort by » oldest newest most voted

answered 2018-05-21 23:26:20 -0600

TomB

1694 ●2 ●11

EnergyPlus has reported a warning, not an error. The warning may not caused directly by your control strategy.

The warning is that alerting you that the chiller performance is being evaluated at a part load condition outside the range described by the performance curve you have specified.

The warning message posted does not, alone, prove the chillers are operating at part loads less than 0.2.

If indeed they are and you want to deactivate the chillers at a given part load ratio, you could use a chiller staging scheme - implemented via a PlantEquipmentOperationCoolingLoad.

edit flag offensive delete link

Comments

@TomB In the timestep with the first error, Chiller Part Load Ratio, Chiller Cycling Ratio and Chiller Electric Energy are all zero, so I guess I can ignore it. I tried setting a gap at zero in PlantEquipmentOperation:CoolingLoad, but that didn't help. I checked the per-timestep outputs and found that the first error occurs at a time when there is no Plant Supply Side Cooling Demand (and water flow rate is zero), but just before a timestep when these values go positive. I'd like to check the other times, but can't see a way to show individual errors instead of the recurring error summary.

Matthias Fripp ( 2018-05-22 20:12:35 -0600 )edit

add a comment

answered 2018-05-22 20:36:00 -0600

rraustad

13804 ●79 ●12 http://www.fsec.ucf.edu/

The chiller is not turning off because the PLR passed to the curve is 0, it is turning off because there is no plant load. The warning is reported when PLR < PLR min, curve so if you decrease PLR min, curve to 0 the warnings will disappear.

For performance curves, the min/max values are meant to limit the curve result so extrapolation does not occur, but what most users do not know/understand is that limiting the curve output can at times cause erroneous results simply because the user did not provide a wide enough range for the curve inputs (e.g., temp or PLR). The best way to avoid issues is to plot the curve across the entire expected range of operation and make sure the curves are robust enough to extrapolate to the extremes of the simulation.

edit flag offensive delete link

Comments

@rraustad I think from the docs that the chiller should run normally for PLR>=0.2 (Minimum Unloading Ratio), use hot-gas bypass for PLR>=0.1 (Minimum Part Load Ratio), and cycle on and off for PLR<=0.1 (which would show a Chiller Cycling Ratio between 0 and 1). For the latter two cases, when the system runs, it should use the amount of power needed for PLR = 0.2, so it never needs to refer to the curve for PLR < 0.2. But I don't know why it doesn't do this for PLR = 0.

Matthias Fripp ( 2018-05-22 20:45:20 -0600 )edit

The amount of power needed at PLR=0 is 0. Consider these equations:

PartLoadRat = 0;
if (PartLoadRat < MinPartLoadRat) FRAC = min(1.0, (PartLoadRat / MinPartLoadRat));
Power = (AvailChillerCap / ReferenceCOP) * ChillerEIRFPLR * ChillerEIRFT * FRAC;
Power = (X) * Y * Z * 0 = 0.

It doesn't matter what EIRfPLR or EIRfT resolve to.

rraustad ( 2018-05-22 21:32:46 -0600 )edit

Unless I'm missing something obvious, I still don't think E+ should issue a warning there.

Julien Marrec ( 2018-05-23 02:49:21 -0600 )edit

Maybe the warning is less meaningful at PLR=0, but as soon as the chiller turned on the warning would be reported (e.g., at PLR = 0.001). What is more meaningful is the recurring warning at the end where the number of times the curve limits are exceeded is shown. Or maybe users should not be notified when performance curves exceed the min/max limits? Developers always struggle with just how much information to give users.

rraustad ( 2018-05-23 07:25:22 -0600 )edit

What @Matthias Fripp in a comment just above is what I think the chiller is doing, and I find his explanation quite clear IMHO. In which case it never needs to refer to a curve below PLR of 0.2, and the curve min Y is 0.18, so I don't understand why the warning is issued here. Are Matthias and I misunderstanding something? Perhaps I should re-read the engineering section...

Julien Marrec ( 2018-05-23 08:49:10 -0600 )edit

add a comment