Space Sizing and HVAC part2 - Code cleanups #10174

mjwitte · 2023-08-25T16:59:09Z

Pull request overview

Builds on Extend Spaces to Sizing and HVAC #9982
Address some past comments in the space heat balance etc world.
This will be a no-diff prep PR before more functional changes in space sizing and HVAC.
Rename variables with redundant/confusing "Zone" in the name, e.g.:

zoneHeatBalance(ZoneNum).ZoneAirHumRat --> zoneHeatBalance.airHumRat
spaceHeatBalance(ZoneNum).ZoneAirHumRat --> spaceHeatBalance.airHumRat

Remove unused variables from DataSingleDuct::AirTerminalMixerData (stumbled across these when rename happened to change ZoneAirHumRat and there were no corresponding changes in SingleDuct.cc.
Move common variables in DataSizing::ZoneSizingData and TermUnitZoneSizingData to a parent struct TermUnitZoneSizingCommonData.

Pull Request Author

Add to this list or remove from it as applicable. This is a simple templated set of guidelines.

Reviewer

This will not be exhaustively relevant to every PR.

…gHVACPart2

mjwitte

Code walkthru. This is a no-diff cleanup PR before more space sizing and HVAC work.
Ready for review @Myoldmopar @rraustad.

mjwitte · 2023-08-29T13:38:02Z

src/EnergyPlus/DataSizing.hh

@@ -339,146 +339,66 @@ namespace DataSizing {
        ZoneSizing zoneSizingMethod = ZoneSizing::Invalid; // load to sizing on: sensible, latent, sensibleandlatent, sensibleonlynolatent
    };

-    struct ZoneSizingData
+    // based on ZoneSizingData but only member variables that are actually used by terminal unit sizing
+    struct TermUnitZoneSizingCommonData


Move common variables in DataSizing::ZoneSizingData and TermUnitZoneSizingData to a parent struct TermUnitZoneSizingCommonData.

mjwitte · 2023-08-29T13:41:56Z

src/EnergyPlus/DataZoneEnergyDemands.hh

-        Real64 ZoneSNLoadPredictedRate = 0.0;                // Predicted sensible load [W] (unmultiplied)
-        Real64 ZoneSNLoadPredictedHSPRate = 0.0;             // Predicted sensible load to heating setpoint [W] (unmultiplied)
-        Real64 ZoneSNLoadPredictedCSPRate = 0.0;             // Predicted sensible load to cooling setpoint [W] (unmultiplied)
-        Real64 ZoneSNLoadHeatRate = 0.0;                     // sensible heating rate [W] (unmultiplied)
-        Real64 ZoneSNLoadCoolRate = 0.0;                     // sensible cooling rate [W] (unmultiplied)
-        Real64 ZoneSNLoadHeatEnergy = 0.0;                   // sensible heating energy [J] (unmultiplied)
-        Real64 ZoneSNLoadCoolEnergy = 0.0;                   // sensible cooling energy [J] (unmultiplied)
+        Real64 predictedRate = 0.0;                          // Predicted sensible load [W] (unmultiplied)
+        Real64 predictedHSPRate = 0.0;                       // Predicted sensible load to heating setpoint [W] (unmultiplied)


Remove Zone from variable names in structs used for both zones and spaces. Redundant and confusing.

mjwitte · 2023-08-29T13:42:47Z

src/EnergyPlus/ZoneTempPredictorCorrector.hh

-        Real64 ZoneTMX = DataHeatBalance::ZoneInitialTemp; // Temporary air temperature to test convergence in Exact and Euler method
-        Real64 ZoneTM2 = DataHeatBalance::ZoneInitialTemp; // Temporary air temperature at timestep t-2 in Exact and Euler method
-        Real64 ZoneT1 = 0.0;                               // Air temperature at the previous time step used in Exact and Euler method
+        Real64 TMX = DataHeatBalance::ZoneInitialTemp; // Temporary air temperature to test convergence in Exact and Euler method
+        Real64 TM2 = DataHeatBalance::ZoneInitialTemp; // Temporary air temperature at timestep t-2 in Exact and Euler method
+        Real64 T1 = 0.0;                               // Air temperature at the previous time step used in Exact and Euler method


More remove Zone from variable names

mjwitte · 2023-08-29T13:51:10Z

src/EnergyPlus/General.hh

+    void MovingAvg(EPVector<Real64> &DataIn, int NumItemsInAvg);
+
    void MovingAvg(Array1D<Real64> &DataIn, int NumItemsInAvg);


@Myoldmopar Some fallout from the consolidation in DataSizing. Some of the common variables were Array1D in ZoneSizingData but were EPVector in TermUnitZoneSizingData. MovingAvg is only applied to the ZoneSizingData side, so the existing function wouldn't accept an EPVector. But MovingAvg is used on a bunch of other Array1Ds so there are now two flavors of MovingAvg. If you prefer I could remove this and change the affected fields in zone sizing back to Array1D.

Myoldmopar

OK, well this is just a whole lot of cleanup. You might have improved our build time just from shorter variable names everywhere lol.

I had one comment about the new almost-duplicate function you had to add. I understand why you added it, and it is certainly possible that we should leave it that way. But let's talk about it first.

Myoldmopar · 2023-08-29T15:29:32Z

src/EnergyPlus/General.cc

+        }
+        DataIn(i) /= NumItemsInAvg; // average to smooth over NumItemsInAvg window
+    }
+}


Question 1:
How often is this function being called? Calling allocate a lot doesn't make me feel great. Is it possible to create a scratch vector on state and when you get inside here, just overwrite the indices of interest and ignore the rest of the vector? Then we wouldn't have to continually call allocate each time we get in here? It may not be worth changing to that, but thought it was worth bringing it up.

Question 2:
Can we just make one MovingAvg function that accepts a gsl::span argument? There has to be a nice way to generalize the functionality so that it can operate on either an Array or a vector regardless of the base index. It's just a series of variables from M to N. Here's an example of using span for both array and vector which does not reference the zero index. I get that there are two issues: whether an Objexx::Array1D be passed as a span, and how complex would it be to handle the array index values generically. So this could easily just be pinned for investigation later, or we can talk about it and I can try to hack around and see what I can get done in a few minutes.

I agree, this was a punt just to get things to build. I can fiddle with it and if I hit a roadblock, I'll holler. There's about 55 calls to MovingAvg and they're all in sizing or in output tables for the component loads reports. So it shouldn't be too hard to move to a common function.

There is a faster way to do this algorithmically. One that uses a single loop, not a nested loop. And it's possible that this faster way also allows you to eliminate the memory allocation. I'll leave this as an exercise to the reader.

@amirroth Once a teacher, always a teacher. Working on my assignment.

Ok, so I had to do this to convince myself that it works. You still need a temporary array but that array is of size NumItemsInAvg and so maybe that can be a std::array and NumItemsInAvg can be a function template variable.

The initialization loop only sums NumItemsInAvg - 1 values. In your example it would only sum 3.

Ah, there it is, for (iWindow = 1; iWindow <= NumItemsInAvg - 1;

Does it complicate things or simplify things to note that Array1D objects can be accessed as a zero-based vector using the square brackets? I wonder, if this function was written as zero based using square brackets, could the function just be a template to take in either a vector or an Array1D? This is just conversation, we don't need to dwell on it any longer.

Array1D or EPVector? If you wanted to handle both vector or EPVector you could use a template or span and square brackets.

Array1D or EPVector? If you wanted to handle both vector or EPVector you could use a template or span and square brackets.

Probably, but I don't speak span or template very well. The bigger issue is getting rid of the tiny values that are causing big diffs. I may just revert all of this out of here and put it in a new branch so the space stuff can move forward.

…gHVACPart2

mjwitte · 2023-08-31T15:18:54Z

src/EnergyPlus/General.cc

+    //if (std::accumulate(DataIn.begin(), DataIn.end(), 0.0) == 0.0) {
+    //    return; // no noeed to average if all zeroes
+    //}


Pushing up the new MovingAvg. The e-17'ish diffs are causing some big diffs in 5ZoneAirCooledWithDOASAirLoop locally (due to x/1e-17) in some system sizing calcs. Thought maybe checking for all zeros and return would help, but the residual value creeps in when there are non-zero values and then the last term gets the tiny value that propagates to the end of the day (where before the last hours of the day were zero).

I'll make a second commit with the zero return active to see if that helps speed-wise - penalty to check, but avoid doing the moving avg on a bunch of zeros.

…gHVACPart2

rraustad

struct reorganiztion and name change only branch, with a couple EPVector swaps

rraustad · 2023-09-13T21:22:50Z

src/EnergyPlus/AirflowNetwork/src/Solver.cpp

-        Real64 CpAir = PsyCpAirFnW(thisZoneHB.ZoneAirHumRat);
-        Real64 RhoAir = PsyRhoAirFnPbTdbW(m_state, m_state.dataEnvrn->OutBaroPress, thisZoneHB.MAT, thisZoneHB.ZoneAirHumRat);
+        Real64 CpAir = PsyCpAirFnW(thisZoneHB.airHumRat);
+        Real64 RhoAir = PsyRhoAirFnPbTdbW(m_state, m_state.dataEnvrn->OutBaroPress, thisZoneHB.MAT, thisZoneHB.airHumRat);
        Real64 InfilVolume = ((exchangeData(ZoneNum).SumMCp + exchangeData(ZoneNum).SumMVCp) / CpAir / RhoAir) * TimeStepSys * Constant::SecInHour;
        Real64 ACH = InfilVolume / (TimeStepSys * m_state.dataHeatBal->Zone(ZoneNum).Volume);


Variable name change only in Solver

rraustad · 2023-09-13T21:23:51Z

src/EnergyPlus/DXCoils.cc

@@ -16370,7 +16370,7 @@ void CalcSecondaryDXCoils(EnergyPlusData &state, int const DXCoilNum)
        } break;
        case CoilDX_MultiSpeedHeating: {
            EvapInletDryBulb = secZoneHB.ZT;
-            EvapInletHumRat = secZoneHB.ZoneAirHumRat;
+            EvapInletHumRat = secZoneHB.airHumRat;
            RhoAir = PsyRhoAirFnPbTdbW(state, state.dataEnvrn->OutBaroPress, EvapInletDryBulb, EvapInletHumRat);
            MSSpeedRatio = thisDXCoil.MSSpeedRatio;
            MSCycRatio = thisDXCoil.MSCycRatio;


Variable name changes only in above 3 source files

rraustad · 2023-09-13T21:28:35Z

src/EnergyPlus/DataSizing.hh

        void scaleZoneCooling(Real64 ratio);
        void scaleZoneHeating(Real64 ratio);
-        void copyFromZoneSizing(ZoneSizingData const &sourceData);
+        void copyFromZoneSizing(DataSizing::ZoneSizingData const &sourceData);
        void allocateMemberArrays(int numOfTimeStepInDay);
    };


good reorg step

rraustad · 2023-09-13T22:11:53Z

src/EnergyPlus/DataZoneEnergyDemands.hh

+        Real64 airSysHeatEnergy = 0.0;                       // air system latent heating energy [J] (unmultiplied)
+        Real64 airSysCoolEnergy = 0.0;                       // latent cooling energy [J] (unmultiplied)
+        Real64 airSysSensibleHeatRatio = 0.0;                // air system SHR []
+        Real64 vaporPressureDifference = 0.0;                // vapor pressure depression [Pa]


Name change only. I see 7 variables now with the same name used for moisture and sensible. I guess that could be a new base class? Not here, just sayin.

rraustad · 2023-09-13T22:15:58Z

src/EnergyPlus/HVACManager.cc

-                    znAirRpt.OABalanceLatentGain = thisZoneHB.MDotOA * (state.dataEnvrn->OutHumRat - thisZoneHB.ZoneAirHumRat) * H2OHtOfVap *
-                                                   TimeStepSysSec * ADSCorrectionFactor;
+                    znAirRpt.OABalanceLatentGain =
+                        thisZoneHB.MDotOA * (state.dataEnvrn->OutHumRat - thisZoneHB.airHumRat) * H2OHtOfVap * TimeStepSysSec * ADSCorrectionFactor;


I see 5 TimeStepSysSec * ADSCorrectionFactor. Could be a local, or does the compiler already basically take care of that? Not here, just sayin.

rraustad · 2023-09-13T22:19:24Z

src/EnergyPlus/SingleDuct.hh

-        this->QZoneMaxRHTempLimitSDAT = 0.0;           // maximum zone heat addition rate given constraints of MaxHeatTemp and max
-        this->MinMassAirFlowSDAT = 0.0;                // the air flow rate during heating for normal acting damper
-        this->QZoneMax2SDAT = 0.0;                     // temporary variable
+        new (this) SingleDuctData();


and I have to remember to use this syntax

rraustad · 2023-09-13T22:20:49Z

src/EnergyPlus/ZoneEquipmentManager.cc

+            for (int i = 0; i < (int)zsFinalSizing.CoolFlowSeq.size(); ++i) {
+                zsFinalSizing.CoolFlowSeq[i] = zsCalcFinalSizing.CoolFlowSeq[i] * TotCoolSizMult;
+                zsFinalSizing.CoolLoadSeq[i] = zsCalcFinalSizing.CoolLoadSeq[i] * TotCoolSizMult;
+            }


EPVector change

rraustad · 2023-09-13T22:21:40Z

src/EnergyPlus/ZoneTempPredictorCorrector.cc

    this->AirPowerCap = 0.0;
-    this->ZoneT1 = 0.0;
+    this->T1 = 0.0;


Name change only

rraustad · 2023-09-13T22:27:10Z

src/EnergyPlus/ZoneTempPredictorCorrector.cc


-    this->ZoneSetPointLast = ZoneSetPoint;
+    this->setPointLast = ZoneSetPoint;
    state.dataHeatBalFanSys->TempZoneThermostatSetPoint(zoneNum) = ZoneSetPoint; // needed to fix Issue # 5048
    state.dataZoneEnergyDemand->DeadBandOrSetback(zoneNum) = thisDeadBandOrSetBack;
    state.dataZoneEnergyDemand->CurDeadBandOrSetback(zoneNum) = thisDeadBandOrSetBack;


Name changes only

rraustad · 2023-09-13T22:30:11Z

tst/EnergyPlus/unit/ZoneTempPredictorCorrector.unit.cc

-    Real64 &SingleZoneCoolRate = thisZoneSysEnergyDemand.ZoneSNLoadPredictedCSPRate;
+    Real64 &SingleZoneTotRate = thisZoneSysEnergyDemand.predictedRate;
+    Real64 &SingleZoneHeatRate = thisZoneSysEnergyDemand.predictedHSPRate;
+    Real64 &SingleZoneCoolRate = thisZoneSysEnergyDemand.predictedCSPRate;
    state->dataHeatBalFanSys->LoadCorrectionFactor.allocate(zoneNum);
    Real64 &CorrectionFactor = state->dataHeatBalFanSys->LoadCorrectionFactor(zoneNum);
    state->dataHeatBal->Zone.allocate(zoneNum);


more name changes through to end

rraustad · 2023-09-13T22:53:43Z

Pulled develop and unit tests run successfully, as expected.

rraustad

This is good to go. Maybe I'll merge later tonight, or let @Myoldmopar have another pass. Since this is just reorg and name changes, I will likely merge tonight.

mjwitte added 6 commits August 24, 2023 12:54

Consolidate ZoneSizingData

cebd1ba

Rename moisture vars with redundant Zone in name

e92e1be

Rename moisture vars with redundant Zone in name - plan B

28d5c4b

Rename sensible vars with redundant Zone in name

b5a2d4f

Merge remote-tracking branch 'remotes/origin/develop' into SpaceSizin…

9658c81

…gHVACPart2

Rename vars in unit tests

6d094b0

mjwitte added the DoNotPublish Includes changes that shouldn't be reported in the changelog label Aug 25, 2023

mjwitte added 2 commits August 28, 2023 15:07

Rename heat balance vars with redundant Zone in name

6037259

Delete unused vars in ATMixer

dfaaacb

mjwitte commented Aug 29, 2023

View reviewed changes

mjwitte marked this pull request as ready for review August 29, 2023 13:52

mjwitte added this to the EnergyPlus 23.2 milestone Aug 29, 2023

Myoldmopar reviewed Aug 29, 2023

View reviewed changes

mjwitte added 4 commits August 29, 2023 10:37

Merge remote-tracking branch 'remotes/origin/develop' into SpaceSizin…

c2b4333

…gHVACPart2

Merge remote-tracking branch 'remotes/origin/develop' into SpaceSizin…

8072205

…gHVACPart2

Revert some EPVectors

e518d22

New MovingAvg and remove duplicate

7a0a420

mjwitte commented Aug 31, 2023

View reviewed changes

mjwitte added 3 commits August 31, 2023 11:40

Add zero check to MovingAvg

0049fda

Merge remote-tracking branch 'remotes/origin/develop' into SpaceSizin…

22cbe94

…gHVACPart2

Revert MovingAvg

ae38207

mjwitte mentioned this pull request Sep 13, 2023

Space Sizing and HVAC Part 3 #10205

Merged

23 tasks

Myoldmopar self-assigned this Sep 13, 2023

rraustad reviewed Sep 13, 2023

View reviewed changes

rraustad approved these changes Sep 13, 2023

View reviewed changes

rraustad merged commit ba46616 into develop Sep 14, 2023
10 checks passed

rraustad deleted the SpaceSizingHVACPart2 branch September 14, 2023 01:32

mjwitte mentioned this pull request Sep 22, 2023

Space Sizing and HVAC Documentation and Unit Test Updates #10239

Merged

26 tasks

mjwitte mentioned this pull request Jun 17, 2024

Space Sizing and HVAC Part 4 #10566

Merged

25 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Space Sizing and HVAC part2 - Code cleanups #10174

Space Sizing and HVAC part2 - Code cleanups #10174

mjwitte commented Aug 25, 2023 •

edited

Loading

mjwitte left a comment

mjwitte Aug 29, 2023

Myoldmopar Aug 29, 2023

mjwitte Aug 29, 2023

Myoldmopar Aug 29, 2023

mjwitte Aug 29, 2023

Myoldmopar Aug 29, 2023

mjwitte Aug 29, 2023

Myoldmopar left a comment

Myoldmopar Aug 29, 2023 •

edited

Loading

mjwitte Aug 29, 2023

amirroth Aug 29, 2023

mjwitte Aug 29, 2023

amirroth Aug 29, 2023

amirroth Aug 31, 2023

rraustad Aug 31, 2023

Myoldmopar Aug 31, 2023

amirroth Aug 31, 2023

mjwitte Aug 31, 2023

mjwitte Aug 31, 2023 •

edited

Loading

rraustad left a comment

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad Sep 13, 2023

rraustad commented Sep 13, 2023

rraustad left a comment

		void MovingAvg(EPVector<Real64> &DataIn, int NumItemsInAvg);

		void MovingAvg(Array1D<Real64> &DataIn, int NumItemsInAvg);

Space Sizing and HVAC part2 - Code cleanups #10174

Space Sizing and HVAC part2 - Code cleanups #10174

Conversation

mjwitte commented Aug 25, 2023 • edited Loading

Pull request overview

Pull Request Author

Reviewer

mjwitte left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Myoldmopar left a comment

Choose a reason for hiding this comment

Myoldmopar Aug 29, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mjwitte Aug 31, 2023 • edited Loading

Choose a reason for hiding this comment

rraustad left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rraustad commented Sep 13, 2023

rraustad left a comment

Choose a reason for hiding this comment

mjwitte commented Aug 25, 2023 •

edited

Loading

Myoldmopar Aug 29, 2023 •

edited

Loading

mjwitte Aug 31, 2023 •

edited

Loading