`Coil:Cooling:DX` cannot mimic `Coil:Cooling:DX:MultiSpeed` PLF > 1 behavior

[shorowit]

Issue overview

I am comparing two models that are identical except one uses Coil:Cooling:DX:MultiSpeed and one uses Coil:Cooling:DX. The results are perfectly identical for many hours, but they diverge whenever the load exceeds the capacity (RTF=1). During those timesteps, they agree on the load, but the Coil:Cooling:DX shows higher electricity use than the Coil:Cooling:DX:MultiSpeed model. I interpret this to mean that the energy reported by the Coil:Cooling:DX is actually using a RTF > 1 instead of limiting it to 1.

Example output:

Top graph is Cooling Coil Runtime Fraction. Bottom graph is Cooling Coil Electricity Energy (J). The bottom graph shows that the models agree when RTF < 1 but diverge when RTF=1.

Operating System (Multiple choices)

Ubuntu

Operating System Version

24.04

Version of EnergyPlus

25.1

Unmethours link or helpdesk ticket number

N/A

Defect file

2 IDFs and EPW:
coildx.zip

[rraustad]

These results are remarkably identical. If I diff the two coil model results they are as you say "perfectly identical". I still can't identify why the electricity consumption power/energy is different. I have ruled out RTF > 1 as the issue.

[rraustad]

The difference is explained by the input for whether to apply the PLF curve to speeds greater than 1. Changing this to yes will show the same power/enrgy. This input is used to disregard startup losses for a compressor that is already running and simply changing capacity. The other coil models do not have this input (would default to Yes in that case).

Coil:Cooling:DX:MultiSpeed,
  mini split heat pump clg coil,          !- Name
  No,                                     !- Apply Part Load Fraction to Speeds Greater than 1

The code in question uses PLR/SpeedRatio if this input is set to No and RTF if set to Yes:

DXCoils::CalcMultiSpeedDXCoilCooling (and Heating):

if (!thisDXCoil.PLRImpact) {
    thisDXCoil.ElecCoolingPower = SpeedRatio * HSElecCoolingPower + (1.0 - SpeedRatio) * LSElecCoolingPower;
} else {
    thisDXCoil.ElecCoolingPower =
        thisDXCoil.CoolingCoilRuntimeFraction * HSElecCoolingPower + (1.0 - thisDXCoil.CoolingCoilRuntimeFraction) * LSElecHeatingPower;
}

[shorowit]

Thanks, @rraustad! I would never have figured that out.

Do you know why the other coil objects do not have this input (or don't assume the behavior equivalent to 'No')? Was it intentional, or any accident? It feels like a deficiency that the other coil objects cannot mimic the default behavior of the multispeed coil.