COPs too high? -- All chilllers? (CR #5781)

[axelstudios]

Added on 2003-05-02 07:56 by @lklawrie

Description

Email from Fred B after brief look at the problem:

I tried a couple of other electric chiller runs but I'm still
getting COPs of 10. The COP seems higher at lower load. I'm
attaching the idf file: 5ZoneWarmest. You might want to send it
to Rich Liesen. The high COPs are evident on the mtr file and
on the csv file.

I ran an engine chiller on the same building. In this case the
COPs seem low: 1.3 to 1.5; I guess I would expect nearly 2. We'll
worry about it later.

I didn't spot the problem in a quick look at the code. Rich may
be able to see what's wrong quickly.

Oh - there is another chiller - the constant COP chiller. But
I have never added autosizing to it. Plus we don't know if it
would work any better.

Analysis from Don Shirey:
I looked at the 5ZoneWarmest file. I see extremely large COPs when I run a summer design day, ranging from 45 in the morning to just over 8 in the heat of the day. The nominal chiller COP is 3.2.

Assuming the equations in SUBROUTINE CalcElectricChillerModel are correct, the curve coefficients may be wrong. The coefficients are (same for this example file and the EngineDrivenChiller example file):

35.0,  !- Temp Design Condenser Inlet {C}
2.778,  !- Temp Rise Coefficient
6.67,  !- Temp Design Evaporator Outlet {C}
autosize,  !- Design Evaporator Volumetric Water Flow Rate {m3/s}
autosize,  !- Design Condenser Volumetric Water Flow Rate {m3/s}
0.9949,  !- Coefficient1 of the capacity ratio curve
-0.045954,  !- Coefficient2 of the capacity ratio curve
-0.0013543,  !- Coefficient3 of the capacity ratio curve
2.333,  !- Coefficient1 of the power ratio curve
-1.975,  !- Coefficient2 of the power ratio curve
0.6121,  !- Coefficient3 of the power ratio curve
0.03303,  !- Coefficient1 of the full load ratio curve
0.6852,  !- Coefficient2 of the full load ratio curve
0.2818,  !- Coefficient3 of the full load ratio curve

In SUBROUTINE CalcElectricChillerModel, the following equations are pertinent:
(starting at line 834 in file PlantChillers.f90)
!Calculate chiller performance from this set of performance equations.
! from BLAST...Z=(TECONDW-ADJTC(1))/ADJTC(2)-(TLCHLRW-ADJTC(3))
DeltaTemp= (TempCondIn - TempCondInDesign) / TempRiseRat &
- (TempEvapOut - TempEvapOutDesign)

      !  from BLAST...RCAV=RCAVC(1)+RCAVC(2)*Z+RCAVC(3)*Z**2
        AvailNomCapRat =   CapacityRat(1)                   &
                         + CapacityRat(2) * DeltaTemp       &
                         + CapacityRat(3) * DeltaTemp ** 2

        AvailChillerCap = ChillerNomCap*AvailNomCapRat

       ! from BLAST...G=ADJEC(1)+ADJEC(2)*RCAV+ADJEC(3)*RCAV**2.
        FullLoadPowerRat=   PowerRat(1)                         &
                          + PowerRat(2) * AvailNomCapRat      &
                          + PowerRat(3) * AvailNomCapRat ** 2

      !  from BLAST...RCLOAD=AMAX1(MINCHFR(I,IPLCTR),AMIN1(CHLRLOAD(I)/CHLROCAP(I) &
      !         /RCAV,MAXCHFR(I,IPLCTR)))
        PartLoadRat = MAX(MinPartLoadRat, MIN(MyLoad/AvailChillerCap,MaxPartLoadRat))

       ! from BLAST...RPOWER=RPWRC(1)+RPWRC(2)*RCLOAD+RPWRC(3)*RCLOAD**2
        FracFullLoadPower = FullLoadFactor(1)                      &
                          + FullLoadFactor(2) * PartLoadRat      &
                          + FullLoadFactor(3) * PartLoadRat ** 2

      ! If FlowLock is True, the new resolved mdot is used to update Power, QEvap, Qcond, and
      ! condenser side outlet temperature.
IF (FlowLock==0) THEN
     QEvaporator = AvailChillerCap * PartLoadRat
     Power = FracFullLoadPower * FullLoadPowerRat * QEvaporator/COP

For (FlowLock == 1) things are calculated slightly differently.
But anyway, for a typical morning hour the condenser water is coming back fairly cool (about 21C).
Following through the calculations:

DeltaTemp = -5.55 C
AvailNomCapRat = 1.2083
AvailChillerCap = 36976
FullLoadPowerRat = 0.84
PartLoadRat = 0.203
FracFullLoadPower = 0.184

Qevaporator = 7518
Power = 363

So, COP = Qevaporator / Power = 20.7!!!!

Qevaporator of 7518 matches MyLoad, which is the load to be met. However, the POWER equation is either wrong or the input coefficients to the various equations are wrong.

And further analysis (Don Shirey)
OK, I believe the equation is wrong. We had a similar issue with DX coil a while back, which is why I changed the part-load input curve for the DX coil from the original DOE-2 method.

The curve FracFullLoadPower already takes into account the part load of the chiller (i.e. the ratio of actual load to full load capacity).

In the POWER calculation, you need to multiply by AvailNomCap instead of QEvaporator. Basically, we're accounting for part load ratio (runtime) twice: once by multiplying AvailNomCap by PartLoadRat, and a second time in FracFullLoadPower which already accounts for PartLoadRat.

This is being done in all chillers...................

When I make the change that I described below, I see COPs of 3.3 at low part-loads to 4.3 on the summer design day. I'm not sure its the "FINAL" answer (maybe there are other problems lurking), but it fixes a major issue........

MJW 25Aug2003
The major issue related to chiller COP is fixed, but there still seems to be a problem with the example files producing COPs consistently higher than nominal. See CR 5945 for followup.

Fix

See description for analysis and fix, also shown below.

Not hearing back from anyone overnight, and spending another few hours on this, I've convinced myself that I've made the correct changes to all chillers except ConstantCOP (correction is not applicable to ConstantCOP chiller). However, someone familiar with the BLAST chiller models needs to check the original BLAST code/documentation to confirm what I have changed.
Originally method:

     QEvaporator = AvailChillerCap * PartLoadRat 
     Power = FracFullLoadPower * FullLoadPowerRat * QEvaporator/COP 

New method:

     QEvaporator = AvailChillerCap * PartLoadRat 
     Power = FracFullLoadPower * FullLoadPowerRat * AvailChillerCap/COP 

For the test file 5ZoneWarmest run on a summer design day with nominal chiller COP of 3.2, I'm now getting COPs of 3.3 to 4.3 (instead of 8 to 45). The new higher COPs (higher than the nominal 3.2) are due to low chiller condenser water inlet temps (21C vs 35C design), and also high chiller evaporator water inlet temps (20-22C vs a more reasonable 15C). The high chiller evap water inlet temps still both me --> yes this is a "variable flow" chiller with "intermittent" pump, but the inlet water temps seem unreasonably high (Rich L --> you may want to take another look at this).

External Ref: Steve Konopacki, SJKonopacki@lbl.gov -- started the ball rolling
Addressed in 03.05.01 V1.1.0.019 by @lklawrie
Last build tested: 03.08.22 V1.1.1.007