Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page

# VRF with constant curves has COP < 1.0

!!!THIS ISSUE SHOULD BE FIXED IN ENERGYPLUS 9.3!!!

I took the DOAtoVRF.idf example file and modified all the curves to be constant, then graphed the COP for the design days and two days in each season. Since the curves are constant and no longer depend on anything, why is the COP changing so much in the simulation?

The COPs are also way too low for heating and cooling, which are 3.5484 and 3.2917 in the IDF, but less than 1.0 and 2.5 in the simulation.

What inputs are left in the VRF object that can be changed to get a VRF system in which there is zero dependence on indoor and outdoor conditions? In other words, how can I get the COP to be a step function?

Here is a google drive link for the IDF and Excel graphs. I discovered this when comparing the energy use of an AC/Furnace, HeatPump, and VRF system for the same building. The energy use for heating and cooling make sense for the AC/Furnace and HeatPump cases, but the VRF system uses about 4/3 as much energy for cooling and 3 times as much energy for heating compared to the HeatPump case, and the COP of the VRF is higher than the HeatPump! This is using the EnergyPlus example VRF curves as well as constant curves. Something is super duper wrong with my IDF, and I've run out of ideas on what could be wrong.

AirConditioner:VariableRefrigerantFlow,
VRF Heat Pump,           !- Heat Pump Name
VRFCondAvailSched,       !- Availability Schedule Name
autosize,                !- Gross Rated Total Cooling Capacity {W}
3.2917,                  !- Gross Rated Cooling COP {W/W}
-5,                      !- Minimum Outdoor Temperature in Cooling Mode {C}
43,                      !- Maximum Outdoor Temperature in Cooling Mode {C}
VRFCoolCapFT,            !- Cooling Capacity Ratio Modifier Function of Low Temperature Curve Name
VRFCoolCapFTBoundary,    !- Cooling Capacity Ratio Boundary Curve Name
VRFCoolCapFTHi,          !- Cooling Capacity Ratio Modifier Function of High Temperature Curve Name
VRFCoolEIRFT,            !- Cooling Energy Input Ratio Modifier Function of Low Temperature Curve Name
VRFCoolEIRFTBoundary,    !- Cooling Energy Input Ratio Boundary Curve Name
VRFCoolEIRFTHi,          !- Cooling Energy Input Ratio Modifier Function of High Temperature Curve Name
CoolingEIRLowPLR,        !- Cooling Energy Input Ratio Modifier Function of Low Part-Load Ratio Curve Name
CoolingEIRHiPLR,         !- Cooling Energy Input Ratio Modifier Function of High Part-Load Ratio Curve Name
CoolingCombRatio,        !- Cooling Combination Ratio Correction Factor Curve Name
VRFCPLFFPLR,             !- Cooling Part-Load Fraction Correlation Curve Name
autosize,                !- Gross Rated Heating Capacity {W}
,                        !- Rated Heating Capacity Sizing Ratio {W/W}
3.5484,                  !- Gross Rated Heating COP {W/W}
-20,                     !- Minimum Outdoor Temperature in Heating Mode {C}
20,                      !- Maximum Outdoor Temperature in Heating Mode {C}
VRFHeatCapFT,            !- Heating Capacity Ratio Modifier Function of Low Temperature Curve Name
VRFHeatCapFTBoundary,    !- Heating Capacity Ratio Boundary Curve Name
VRFHeatCapFTHi,          !- Heating Capacity Ratio Modifier Function of High Temperature Curve Name
VRFHeatEIRFT,            !- Heating Energy Input Ratio Modifier Function of Low Temperature Curve Name
VRFHeatEIRFTBoundary,    !- Heating Energy Input Ratio Boundary Curve Name
VRFHeatEIRFTHi,          !- Heating Energy Input Ratio Modifier Function of High Temperature Curve Name
WetBulbTemperature,      !- Heating Performance Curve Outdoor Temperature ...
edit retag close merge delete

Sort by ยป oldest newest most voted

I am seeing very similar results in V9.3 (not yet released) so this is an issue in EnergyPlus. I would agree that COP should be constant with these inputs (unless I'm missing something). I will post an issue on GitHub. I also see the VRF HP PLR and RTF are not equal using the unity Cooling/Heating Part-Load Fraction Correlation Curves, this also seems wrong.

Update 1/21/20:

The Cooling/Heating Energy Input Ratio Modifier Function of Low/High Part-Load Ratio Curve cannot be 1. It must be proportional to PLR. When I made those changes in 4 places I get this. I'm still not sure of the changing cooling COP but I am getting closer to what you expect.

Further Update 1/21/20:

The changing cooling COP was due to the Cooling/Heating Part-Load Fraction Correlation Curve not being 1.

Update 1/22/20:

These curves cannot be 1 and must be proportional to PLR (i.e., second term = 1, not the first). Actually, I think the High EIRfPLR curves can remain 1 since when the high PLR curve is used when the PLR >= 1.

CoolingEIRLowPLR, !- Cooling Energy Input Ratio Modifier Function of Low Part-Load Ratio Curve Name
CoolingEIRHiPLR, !- Cooling Energy Input Ratio Modifier Function of High Part-Load Ratio Curve Name
HeatingEIRLowPLR, !- Heating Energy Input Ratio Modifier Function of Low Part-Load Ratio Curve Name
HeatingEIRHiPLR, !- Heating Energy Input Ratio Modifier Function of High Part-Load Ratio Curve Name


And these curves must be 1 so no cycling losses are modeled.

VRFCPLFFPLR,   !- Cooling Part-Load Fraction Correlation Curve Name
VRFCPLFFPLR,   !- Heating Part-Load Fraction Correlation Curve Name

more

( 2020-01-16 13:01:15 -0500 )edit

@rraustad Any recommendations for a workaround to simulating VRF? I'm willing to kluge together some ZoneHVAC objects, ElectricEquipment, and EMS to use the curves in EnergyPlus, or anything better you'd recommend. We have a project using these objects that is due in July 2020. Maybe the bug will be fixed by then and I should just wait? (I have no idea in the level of complication present in debugging the code.)

( 2020-01-21 10:15:34 -0500 )edit
1

@mldichter after making the changes I describe above I get a constant Cooling/Heating COP as you would expect. The operating COP is lower because it includes all power consumed by the system (e.g., condenser power, TU fan power, etc.). Does this help?

( 2020-01-21 14:57:07 -0500 )edit

@rraustad I think so, but the overall COP seems pretty low. To get that close to 0 COP, the compressor would have to be off with fans, etc running, but at the same time the cooling COP is high.

( 2020-01-21 16:31:56 -0500 )edit

@rraustad - I just independently discovered this & a temporary work-around is to set the minimum PLR equal to 1.0. Can you expand on which curves you modified & how?

edit - figured out the curve. For a constant COP, the EIRPLR curve should be:

Curve:Cubic, VRFEIRPLR, 0, !-C1 1, !-C2 0, !-C3 0, !-C4 0, !- minimum X 1.5; !- maximum X

& VRFs can be modeled properly with modified EIRPLR Curves where the result is the EIRPLR * PLR

( 2020-01-21 18:54:07 -0500 )edit