Knowledgebase : Central Plant > Chillers
Dear members of the e+ team:

I'm currently doing a simulation and I find something strange with chiller performance:

EIRFPLR ratio value is 0.59
EIRFTEMP ratio is 0.861

(both taken from the outputs)

reference COP is 3.02 (from the input chiller:electric:EIR)

BUT COP has a value of 2.00 at the same time! (also taken fron the outputs)

Shouldn't it be of about 5,94?

I cannot find the reason. If you could help with it, I'd be grateful.


To answer your question, COP is calculated as Qactual / Pactual. These values are calculated based on performance curves selected by the user. The most confusion lies with the EIRFPLR curve which includes both the variation in COP as a function of PLR AND the fact that power is reduced as PLR is reduced. When creating the EIRFPLR curve coefficients use the following equation:

EIRFPLR = Pactual/(Preference * CapFT * EIRFT)

This appears to be more of a power ratio term instead of an efficiency term (EIR), but the model was developed by others and we simply implemented that model as defined in the literature. See the references for this model for further information.

If existing performance curves are used (see DataSets folder in the E+ installation directory), be sure that these curves represent the selected chiller. It is always good to plot these curves and compare to manufacturers data. Once this comparison is completed, calculating the coefficients is fairly easy.
Question: Please explain how to calculate the Generator Heat Input Function of Part Load Ratio Curve (GenHir curve) for Chiller:Absorption:Indirect.

Answer: Referring to the equations in the Engineering Reference section "Indirect Absorption Chiller" starting on pdf p. 487 (v5.0), with some additional notes:

PLR: The part-load ratio of the indirect absoprtion chiller’s evaporator is simply the actual cooling
effect required (load) divided by the maximum cooling effect available.

PLR = Qevap/Qevapmax

where
PLR = part-load ratio of chiller evaporator
Qevap = chiller evaporator operating capacity [W] (current load on the chiller)
Qevapmax = chiller evaporator available capacity [W] (Nominal capacity times temperature correction curves)

Qgenerator = GeneratorHIR * Qevapmax * GenfCondT * GenfEvapT * CyclingFrac

So, GeneratorHIR = Qgenerator / (Qevapmax * GenfCondT * GenfEvapT * CyclingFrac)

If operating at nominal condenser and evaporator temepratures such that GenfCondT = 1 and GenfEvapT = 1, and if operating above PLRmin such that CyclingFrac = 1, then:

GeneratorHIR = Qgenerator / Qevapmax

Since PLR = Qevap/Qevapmax, then Qevapmax = Qevap/PLR

GeneratorHIR = Qgenerator*PLR/Qevap

or, stated another way GeneratorHIR = PLR/COP

So, GeneratorHIR(fPLR) is *not* normalized by a nominal COP or HIR. When PLR = 1.0, GeneratorHIR = 1/COP.

Also, because the denominator is Qevapmax, GeneratorHIR approaches zero as PLR approaches zero.

For example, for a perfect chiller which has constant efficiency at all PLRs, with a nominal COP of 0.5 at full load at rated temperatures, the GeneratorHIR(fPLR) values would be:

PLR = 0.0, GeneratorHIR = 0.0 (PLR/COP = 0.0/0.5 = 0.0)
PLR = 0.2, GeneratorHIR = 0.4 (PLR/COP = 0.2/0.5 = 0.4)
PLR = 0.5, GeneratorHIR = 1.0 (PLR/COP = 0.5/0.5 = 1.0)
PLR = 1.0, GeneratorHIR = 2.0 (PLR/COP = 1.0/0.5 = 2.0)

The COP of a chiller is a function of part load ratio. It is mainly determined by the Energy Input to Cooling Output Ratio Function of Part Load Ratio Curve. When the EIR model is used for an electric chiller, the curve has an independent variable: part load ratio. For the ReformulatedEIR model, the curve requires two independent variables: leaving condenser water temperature and part load ratio. Each independent variable has its min and max values. If a variable is outside the allowed range, the nearest allowed value is used, possibly resulting in an unexpected result.

If you would like to compare COP values for two types of chillers, you may need to ensure that the same conditions are applied. For simplicity, you may want to use a spreadsheet to calculate the curve values.