HPWH shows very high COP

[rraustad]

User file shows COP = 10 when COP is set at 2.7. For loop in code inadvertently increases tank source node entering water temperature which in turn lowers source node water mass flow rate, which in turn lowers HPWH PLR to the point where COP is high.

2 places in code starting at line 6856 of PlantWaterThermalTanks.cc

for ( loopIter = 1; loopIter <= 4; ++loopIter ) {
CalcHPWHDXCoil( HPWaterHeater( HPNum ).DXCoilNum, HPPartLoadRatio );
// Currently, HPWH heating rate is only a function of inlet evap conditions and air flow rate
// If HPWH is ever allowed to vary fan speed, this next sub should be called.
// (possibly with an iteration loop to converge on a solution)
// CALL CalcDOE2DXCoil(DXCoilNum, HPPartLoadRatio, FirstHVACIteration,PartLoadRatio, FanOpMode)
CondenserDeltaT = Node( HPWaterOutletNode ).Temp - Node( HPWaterInletNode ).Temp;
// move the full load outlet temperature rate to the water heater structure variables
// (water heaters source inlet node temperature/mdot are set in Init, set it here after CalcHPWHDXCoil has been called)
WaterThermalTank( WaterThermalTankNum ).SourceInletTemp = Node( HPWaterInletNode ).Temp + CondenserDeltaT;
WaterThermalTank( WaterThermalTankNum ).SourceMassFlowRate = MdotWater;

//           this CALL does not update node temps, must use WaterThermalTank variables
// select tank type
{ auto const SELECT_CASE_var1( HPWaterHeater( HPNum ).TankTypeNum );
    if ( SELECT_CASE_var1 == MixedWaterHeater ) {
        CalcWaterThermalTankMixed( WaterThermalTankNum );
        NewTankTemp = WaterThermalTank( WaterThermalTankNum ).TankTemp;
    } else if ( SELECT_CASE_var1 == StratifiedWaterHeater ) {
        CalcWaterThermalTankStratified( WaterThermalTankNum );
        NewTankTemp = FindStratifiedTankSensedTemp( WaterThermalTankNum, HPWaterHeater( HPNum ).ControlSensorLocation );
    }}
    Node( HPWaterInletNode ).Temp = WaterThermalTank( WaterThermalTankNum ).SourceOutletTemp;
    if ( std::abs( Node( HPWaterInletNode ).Temp - HPWHCondInletNodeLast ) < SmallTempDiff ) break;
    HPWHCondInletNodeLast = Node( HPWaterInletNode ).Temp;

}

Tests removing the for loop show what appears to be proper operation.
This loop was added for stratified tanks so close review is necessary.
It may also be that this was added to correct a defect file, where tank temperatures may have been changing rapidly and required some iteration. This is a little scary to change except for the fact that the HPWH is obviously not working correctly now.

[mjwitte]

@rraustad Could you keep the for loop but prevent the coil load variable from incrementing and get it to recalc each time through the loop?

[rraustad]

@mjwitte I don't think so, the HPWH coil load is based solely on the coil capacity, the operating PLR, the evaporator air inlet temp, and the HPWH condenser water inlet temp.

This section of the code is intended to provide the maximum available output of the HPWH with PLR=1. What this loop does is model the HPWH then tank. I guess the idea was to iterate on a solution, however, at full capacity, the tank temperature will tend to change. On the second iteration, the new tank temperature, or more specifically, the new tank source node outlet temperature, is used as the starting point for the second iteration. This is incorrect. If the tank temperature from the last time step were used, the same answer would be found, so there is no need for iteration.

What I see is that the HWPH gives the full capacity result, then the tank models the new tank temp given the HPWH (source nodes) input to the tank. For example, a tank at 60C, given some load, models the HPWH and finds a new tank temp of 63C. This loop is using 63C as the input for the next iteration, when it should use 60C. Since the HPWH "load" is based on a set point, this loop has the effect of causing the source node mass flow rate to decrease since the HPWH "sees" a higher inlet water temp. If I change it to use 60C on the second iteration, the new tank temp will again be 63C (since PLR and inlet air and water temps haven't changed) so why iterate?

I know Brent did this for a reason, I just don't know what that reason was.

[rraustad]

I just tried keeping the for loop and not updating the HPWH condenser inlet node based on the new tank temp. The answer from the 1st and 2nd iterations are the same. I do see in the results that the HPWH Cooling Coil Total Water Heating Rate [W] is slightly different from the Water Heater Source Side Heat Transfer Rate [W] and the difference is greater at lower tank temps. Since these should match, maybe that is what Brent was trying to correct?

[mjwitte]

@rraustad I've been looking back through source code and it looks like this particular loop hasn't changed since at least v7.0 maybe even longer. But in the PlantWaterThermalTank.f90 history I see this:

3/15/2012 4:44:12 PM zs_NoCRs_Brent Griffith EnergyPlus allow heat pump water heater to use stratified tank 1.77

That would have been during the v7.1 development cycle. I can send you zips of v7.0 and v7.1 f90 source if you want. Not sure if this is where things went wrong, or one of many other changes to this file since then.

[Myoldmopar]

Wanna see something cool? You can check where each line of a source file came from. If you go to the EnergyPlusArchive repo, browse to the PlantWaterThermalTank.f90 file and click "Blame", it will take you here. You can then go line-by-line and look at when it was changed last, who did it, and what commit it came in with. You could then click on the commit code and see all the changes that occurred at that time.

[mjwitte]

That's useful.

On 9/4/2014 10:17 AM, Edwin Lee wrote:

Wanna see something cool? You can check where each line of a source
file came from. If you go to the EnergyPlusArchive repo, browse to the
PlantWaterThermalTank.f90 file and click "Blame", it will take you
here
https://github.com/NREL/EnergyPlusArchive/blame/master/SourceCode/PlantWaterThermalTank.f90.
You can then go line-by-line and look at when it was changed last, who
did it, and what commit it came in with. You could then click on the
commit code and see all the changes that occurred at that time.

—
Reply to this email directly or view it on GitHub
#4455 (comment).

[rraustad]

Here is the result of the test suite run. Only 1 file had diff's even though there are 2 example files, one mixed and one stratified. In the mixed file, the before and after do not show any better agreement. I think I know why. When the HPWH is tested for it's full capacity (this part I think is OK), and the resulting tank temp is greater than the SP, only the tank source node flow rate is iterated on using RegulaFalsi. I wonder if the coil heating rate and tank heating rate would converge on the same answer if both the coil and tank were simulated inside the iteration loop.

[rraustad]

Given that the tank is simulated on a 1 sec interval and the HPWH is simulated at the time step interval, and that the tank source node calcs are integrated over the simulated time period, can these 2 reports even be equal?

Since I've had no luck correcting the discrepancy between the HPWH and tank source heating rate, I think I'll correct the problem with the additional loop iterations, and leave this issue for another time?

[rraustad]

Committing the fix for high HPWH COP and will create a separate issue regarding differences in HPWH and tank source node heat transfer rates.

[rraustad]

This actually looks better than before. Not sure why.

[rraustad]

I should add that:

HPWH COP = HPWHDXCOIL:Cooling Coil Total Water Heating Rate / HPWHDXCOIL:Cooling Coil Water Heating Electric Power

and

COP based on tank heating rate = J2:Water Heater Source Side Heat Transfer Rate / HPWHDXCOIL:Cooling Coil Water Heating Electric Power