Correct physics based VRF model to align inlet and outlet air flow rates #7295

rraustad · 2019-05-09T18:47:13Z

Pull request overview

The physics based VRF model calculates the operating evaporator and condenser temperatures after all terminal units have been simulated. The method used adjusted the terminal unit simulated inlet node air flow rate such that it did not match the TU outlet node air flow rate. This change fixes #6301.

Work Checklist

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

Title of PR should be user-synopsis style (clearly understandable in a standalone changelog context)
At least one of the following appropriate labels must be added to this PR to be consumed into the changelog:
- Defect: This pull request repairs a github defect issue. The github issue should be referenced in the PR description
- Refactoring: This pull request includes code changes that don't change the functionality of the program, just perform refactoring
- NewFeature: This pull request includes code to add a new feature to EnergyPlus
- Performance: This pull request includes code changes that are directed at improving the runtime performance of EnergyPlus
- DoNoPublish: This pull request includes changes that shouldn't be included in the changelog

Review Checklist

This will not be exhaustively relevant to every PR.

rraustad · 2019-05-09T18:48:20Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

@@ -8663,34 +8663,35 @@ namespace HVACVariableRefrigerantFlow {
        }

        // Set inlet air mass flow rate based on PLR and compressor on/off air flow rates
-        SetAverageAirFlow(VRFTUNum, 1.0, temp);
+        //SetAverageAirFlow(VRFTUNum, 1.0, temp);


The TUs have already been simulated. At this point it appears the model want to know what the evap T and cond T are for use during the next time step. Anything that tried to resimulate the TUs is now commented out.

Note here that the original model did not set OA flow rate and therefore used the OA flow for the last TU simulated when SetAverageAirFlow was called. When I corrected this (lines 8671 and 8676), my new method matched the old method results.

Also note that any calls to this and other functions updates the node flow rates. So data set when the TU was actually modelled is used here instead.

rraustad · 2019-05-09T18:48:40Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc


        // Simulation the OAMixer if there is any
        if (this->OAMixerUsed) {
-            SimOAMixer(this->OAMixerName, false, this->OAMixerIndex);
+            //SimOAMixer(this->OAMixerName, false, this->OAMixerIndex);


Mixer has already been simulated.

rraustad · 2019-05-09T18:49:31Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

-            T_coil_in = Node(FanOutletNode).Temp;
-            W_coil_in = Node(FanOutletNode).HumRat;
-        }
+        //if (this->FanPlace == BlowThru) {


Fan has already been simulated. If the intent here is to simulate at full flow then why weren't the coils simulated again?

rraustad · 2019-05-09T18:51:43Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

@@ -8702,8 +8703,9 @@ namespace HVACVariableRefrigerantFlow {
        if ((Garate > 0.0) && ((!VRF(VRFNum).HeatRecoveryUsed && CoolingLoad(VRFNum)) ||
                               (VRF(VRFNum).HeatRecoveryUsed && TerminalUnitList(TUListIndex).HRCoolRequest(IndexToTUInTUList)))) {
            // 1.1) Cooling coil is running
-            QZnReqSenCoolingLoad = max(0.0, -1.0 * ZoneSysEnergyDemand(ZoneIndex).OutputRequiredToCoolingSP);
-            Tout = T_TU_in - QZnReqSenCoolingLoad * 1.2 / Garate / 1005;
+            //QZnReqSenCoolingLoad = max(0.0, -1.0 * ZoneSysEnergyDemand(ZoneIndex).OutputRequiredToCoolingSP);


Here's the big question. What Tout is needed here? The Tout when the coil is operating I assume. Line 8706 is wrong since for that zone, the TU operated and probably met the load, so OutputRequiredToCoolingSP is probably 0. Instead I use the actual coil outlet temp, but this is premised on needing actual Tout.

I guess I should read the Eng. Ref. to get better acquainted.
@nealkruis @mbadams5 @mjwitte @Myoldmopar @jasondegraw @kbenne this will need some serious thought and review because I don't know this model. What the heck should it do? Does it need full output conditions while coils are operating? Yet the coils can have different PLR's so how to average correctly to get condenser performance? Anyone recognize old line 8706 above? Q*1.2/Mdot/1005 ? Figuring out what this function needs will probably correct this issue. If Tout at full air flow is really needed then the coils will need to be simulated again, or better yet that information needs to be saved when the TUs model full flow and then used here. Is Xiufeng Pang or RP Zhang or anyone at LBNL available for consult on this model?

…unit-flow-balance-problems

…1-VRF-Terminal-unit-flow-balance-problems

…//github.com/NREL/EnergyPlus into 6301-VRF-Terminal-unit-flow-balance-problems

rraustad · 2019-05-14T15:53:18Z

Since the answers didn't line up well with develop I decided to mimic what the original model does. The answers don't line up perfectly because there was a mistake in the original code. I also don't understand why the model assumes full flow rate when determining suction/discharge temperature at the condenser when actual TU flow rates used are known and coil outlet temps are also known. Regardless, these changes mimic the original model.

Air flow rate is now consistent and zone temperature is roughly unchanged. Develop is left column, this branch is right column. In this first figure since zone temperature is held to set point, the TU outlet node flow rate was correct so in the right column, the new flows should roughly match the blue line in the left figure.

The compressor speed and cooling coil rate are also roughly the same.

The evap/condenser temps used are also similar but won't match because of the problem in code. Where full flow was set, the developer forgot to set the OA flow rate and therefore when SetAverageAirFlow was called it used the OA flow rate from the last TU simulated and wound up with a different mixed air temp.

rraustad · 2019-05-14T15:57:59Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

            VRFTU(VRFTUNum).ControlVRF_FluidTCtrl(VRFTUNum, QZnReq, FirstHVACIteration, PartLoadRatio, OnOffAirFlowRatio);
            VRFTU(VRFTUNum).CalcVRF_FluidTCtrl(VRFTUNum, FirstHVACIteration, PartLoadRatio, SysOutputProvided, OnOffAirFlowRatio, LatOutputProvided);
+            if ( PartLoadRatio == 0.0 ) { // set coil inlet conditions when coil does not operate. Inlet conditions are set in ControlVRF_FluidTCtrl when PLR=1


Pick up temps used to calculate evap/cond temps later on. Here only for PLR=0.

rraustad · 2019-05-14T16:00:12Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

-                                TotalTUHeatingCapacity,
-                                TerminalUnitList(TUListNum).TotalHeatLoad,
-                                MaxHeatingCapacity(VRFCond));
+TerminalUnitList(TUListNum).TotalCoolLoad,


Oops, will fix this on next commit.

rraustad · 2019-05-14T16:01:29Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

@@ -10186,6 +10202,14 @@ namespace HVACVariableRefrigerantFlow {
        // Otherwise the coil needs to turn on. Get full load result
        PartLoadRatio = 1.0;
        this->CalcVRF_FluidTCtrl(VRFTUNum, FirstHVACIteration, PartLoadRatio, FullOutput, OnOffAirFlowRatio);
+        this->fanOutletT = Node(this->fanOutletNode).Temp;


Here, when the TU operates the conditions resulting from PLR=1 (full flow to mimic what was done before) are captured and used in calculations for evap/cond temps.

rraustad · 2019-05-14T16:02:27Z

src/EnergyPlus/HVACVariableRefrigerantFlow.hh

@@ -700,6 +700,10 @@ namespace HVACVariableRefrigerantFlow {
        int ATMixerSecNode;              // secondary air inlet node number for the air terminal mixer
        int ATMixerOutNode;              // outlet air node number for the air terminal mixer
        bool firstPass;                  // used to reset global sizing data
+        Real64 coilInNodeT;              // coil inlet node temp at full flow (C)


new variables to hold data in order to eliminate the problem with flow mismatch from inlet to outlet.

rraustad · 2019-05-15T14:24:10Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

@@ -3287,6 +3287,8 @@ namespace HVACVariableRefrigerantFlow {
                    if (errFlag) {
                        ShowContinueError("...occurs in " + cCurrentModuleObject + " = " + VRFTU(VRFTUNum).Name);
                        ErrorsFound = true;
+                    } else {
+                        VRFTU(VRFTUNum).fanOutletNode = Fans::Fan(VRFTU(VRFTUNum).FanIndex).OutletNodeNum;


Save fan outlet node for use in TeTc calculations.

Seems reasonable enough, if you are going to have to query what the fan is doing, you can check the outlet node.

rraustad · 2019-05-15T17:54:27Z

I should probably show the marked difference between this branch and develop. Since the original model tried to use full flow in the TeTc calc's, and forgot to also set OACompOnMassFlow along with CompOnMassFlow, the resulting TeTc is now different. This impacts the air flow rate and will show diff's. Other than this, right or wrong the model still does what it originally did.

rraustad · 2019-05-15T17:55:38Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

-            W_coil_in = Node(FanOutletNode).HumRat;
-        }
+        T_coil_in = this->coilInNodeT;
+        W_coil_in = this->coilInNodeW;


Change here is that coil inlet temps are now used so it doesn't matter if blow thru fan or OA mixer is used. Gets rid of a lot of unnecessary code.

👍 👍 👍 👍 👍 👍 👍 👍 👍 👍 👍

rraustad · 2019-05-16T19:46:50Z

The 2 new diff's are from the merge of #7258 into develop. No need to spin off CI again just for that.

Myoldmopar · 2019-05-17T01:16:28Z

But there are build warnings. It looks like just one unused "using" statement (which is interpreted as an unused variable in a function). The error message is difficult to parse, so the CI message doesn't always point you to the right spot easily, but I can track it down for you if you need help diagnosing where it is.

Myoldmopar · 2019-05-17T02:04:12Z

...ce/src/simulation-models-encyclopedic-reference-002/variable-refrigerant-flow-heat-pumps.tex

@@ -895,7 +895,7 @@ \subsubsection{Overview}\label{VRF-FluidTCtrl-HP-overview}

 Note that a number of calculation steps are coupled together in the VRF-FluidTCtrl model, for instance, the piping loss calculation and the system performance calculation. More specifically, the piping loss changes the operating conditions of the system, which may lead to a different control strategy and thus in reverse affect the amount of piping loss. This makes it difficult to obtain an analytical solution for a number of operational parameters (e.g., enthalpy of refrigerant entering the indoor unit), and therefore numerical iterations are employed to address this problem (refer to Figure VRF-FluidTCtrl-3 for more details). Therefore, the VRF-FluidTCtrl model can have a longer execution time to perform the simulation than the VRF-SysCurve model. 

-The object connections for VRF-FluidTCtrl model is similar to those for VRF-SysCurve model. The difference lies in the object used to describe a specific components. For example, VRF-SysCurve model uses \emph{AirConditioner:VariableRefrigerantFlow} object to describe the VRF outdoor unit performance, while in VRF-FluidTCtrl model the \emph{AirConditioner:VariableRefrigerantFlow} object is used.
+The object connections for VRF-FluidTCtrl model is similar to those for VRF-SysCurve model. The difference lies in the object used to describe a specific components. For example, VRF-SysCurve model uses \emph{AirConditioner:VariableRefrigerantFlow} object to describe the VRF outdoor unit performance, while in VRF-FluidTCtrl model the \emph{AirConditioner:VariableRefrigerantFlow:FluidTemperatureControl} object is used.


Myoldmopar · 2019-05-17T02:04:51Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

@@ -3287,6 +3287,8 @@ namespace HVACVariableRefrigerantFlow {
                    if (errFlag) {
                        ShowContinueError("...occurs in " + cCurrentModuleObject + " = " + VRFTU(VRFTUNum).Name);
                        ErrorsFound = true;
+                    } else {
+                        VRFTU(VRFTUNum).fanOutletNode = Fans::Fan(VRFTU(VRFTUNum).FanIndex).OutletNodeNum;


Seems reasonable enough, if you are going to have to query what the fan is doing, you can check the outlet node.

Myoldmopar · 2019-05-17T02:06:42Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

@@ -8526,7 +8537,7 @@ namespace HVACVariableRefrigerantFlow {
            for (int i = 1; i <= TerminalUnitList(TUListNum).NumTUInList; i++) {
                int VRFTUNum = TerminalUnitList(TUListNum).ZoneTUPtr(i);
                // analyze the conditions of each IU
-                VRFTU(VRFTUNum).CalcVRFIUVariableTeTc(VRFTUNum, EvapTemp(i), CondTemp(i));
+                VRFTU(VRFTUNum).CalcVRFIUVariableTeTc(EvapTemp(i), CondTemp(i));


👍 👍 👍 👍

Myoldmopar · 2019-05-17T02:07:10Z

src/EnergyPlus/HVACVariableRefrigerantFlow.cc

-            W_coil_in = Node(FanOutletNode).HumRat;
-        }
+        T_coil_in = this->coilInNodeT;
+        W_coil_in = this->coilInNodeW;


👍 👍 👍 👍 👍 👍 👍 👍 👍 👍 👍

Myoldmopar · 2019-05-17T02:08:05Z

tst/EnergyPlus/unit/HVACVariableRefrigerantFlow.unit.cc

+    EXPECT_EQ(VRF(IndexVRFCondenser).IUEvaporatingTemp, 15);
+    // fan heat added to coil inlet temp, coil surface temp should also be lower to account for less heat tranfer
+    EXPECT_LT(VRF(IndexVRFCondenser).IUCondensingTemp, saveHeatingCondTemp);
+


Great unit test.

Myoldmopar · 2019-05-17T02:08:33Z

tst/EnergyPlus/unit/HVACVariableRefrigerantFlow.unit.cc

+    EXPECT_EQ(VRF(IndexVRFCondenser).IUEvaporatingTemp, 15);
+    // fan heat added to coil inlet temp, coil surface temp should also be lower to account for less heat tranfer
+    EXPECT_LT(VRF(IndexVRFCondenser).IUCondensingTemp, saveHeatingCondTemp);
+
    // Clean up
    VRF.deallocate();


Ideally these are cleaned up as part of the clear_state after every single unit test is run, but I understand it doesn't hurt to have them here.

…unit-flow-balance-problems

Myoldmopar · 2019-05-17T02:16:55Z

Alright, I fixed that unused variable issue, merged up from develop, and will let this run over the next couple hours. I'm anticipating very clean CI except for the actual expected diff file. If so I'll merge in the morning.

rraustad · 2019-05-17T02:21:40Z

This:

// Clean up
VRF.deallocate();

should have been removed. I was so focused on testing this that I missed this. We should not allow deallocate in unit tests? New python script warranted?

rraustad · 2019-05-17T02:27:10Z

And this:

made me think it was something other than what I worked on (why would I think that if this warning never showed up before?). Thanks for catching that.

Myoldmopar · 2019-05-17T02:27:26Z

New python script

I was already half asleep and this phrase woke me right back up. I like your thinking but I would worry that in this particular case, it may hit too many false alarms. There are certainly times where you want to deallocate/reallocate within a unit test, and I feel like we may end up hitting too many of those cases to warrant a script. But your line of thinking here makes me smile.

Myoldmopar · 2019-05-17T02:29:34Z

Yeah, the full warning message text is:

$ make HVACVariableRefrigerantFlow.cc.o
Building CXX object src/EnergyPlus/CMakeFiles/energypluslib.dir/HVACVariableRefrigerantFlow.cc.o
In file included from /eplus/repos/4eplus/src/EnergyPlus/HVACVariableRefrigerantFlow.cc:86:
/eplus/repos/4eplus/src/EnergyPlus/MixedAir.hh: In member function ‘void EnergyPlus::HVACVariableRefrigerantFlow::VRFTerminalUnitEquipment::CalcVRFIUVariableTeTc(Real64&, Real64&)’:
/eplus/repos/4eplus/src/EnergyPlus/MixedAir.hh:414:34: warning: unused variable ‘EnergyPlus::MixedAir::OAMixer’ [-Wunused-variable]
     extern Array1D<OAMixerProps> OAMixer;
                                  ^~~~~~~

Which makes it pretty clear in the third or so line that the error occurs in In member function ‘void EnergyPlus::HVACVariableRefrigerantFlow::VRFTerminalUnitEquipment::CalcVRFIUVariableTeTc. However, Decent CI has trouble with parsing messages that run this long and detecting which part of it is the actual useful error content. It's easy enough for me to build the code and find it locally.

rraustad · 2019-05-17T02:29:35Z

Shoot, now I have to think about a valid deallocate in a unit test...

Myoldmopar · 2019-05-17T02:31:56Z

Shoot, now I have to think about a valid deallocate in a unit test...

Yeah, maybe there's not one, I may be concerned at that possibility for nothing. For better or worse (worse) there definitely are a ton of deallocate calls in tst/EnergyPlus/unit though.

rraustad · 2019-05-17T02:33:52Z

"definitely are a ton of deallocate calls in tst/EnergyPlus/unit"

Of course there are, we are old and lazy!!!

Myoldmopar · 2019-05-17T11:48:54Z

Everything looks great. Merging.

Adjust cryptic model to align inlet/outlet air flow rates

3aa02d6

rraustad added the Defect Includes code to repair a defect in EnergyPlus label May 9, 2019

rraustad commented May 9, 2019

View reviewed changes

Myoldmopar and others added 5 commits May 13, 2019 09:37

Merge remote-tracking branch 'origin/develop' into 6301-VRF-Terminal-…

63442e6

…unit-flow-balance-problems

Add a few comments

03f4064

Mimic new VRF model calculations

ac463cb

Merge branch 'develop' of https://github.com/NREL/EnergyPlus into 630…

a674053

…1-VRF-Terminal-unit-flow-balance-problems

Merge branch '6301-VRF-Terminal-unit-flow-balance-problems' of https:…

1b8245c

…//github.com/NREL/EnergyPlus into 6301-VRF-Terminal-unit-flow-balance-problems

rraustad commented May 14, 2019

View reviewed changes

rraustad changed the title ~~Adjust cryptic model to align inlet/outlet air flow rates~~ Correct physics based VRF model to align inlet and outlet air flow rates May 14, 2019

rraustad commented May 15, 2019

View reviewed changes

rraustad added 3 commits May 15, 2019 13:20

Clean up and add unit test

69211ef

A bit more clean up to fix formatting and delete unnecessary code

ce7d093

Delete unused variable

b2f4106

rraustad commented May 15, 2019

View reviewed changes

Remove more unused variables

5bd6ae1

Myoldmopar approved these changes May 17, 2019

View reviewed changes

Myoldmopar added 2 commits May 16, 2019 21:12

Merge remote-tracking branch 'origin/develop' into 6301-VRF-Terminal-…

2ec3bcf

…unit-flow-balance-problems

Fix unused variable issue

9792e73

Myoldmopar merged commit c94a1d3 into develop May 17, 2019

Myoldmopar deleted the 6301-VRF-Terminal-unit-flow-balance-problems branch May 17, 2019 11:49

rraustad mentioned this pull request Feb 11, 2020

Allow ZoneHVAC:TerminalUnit:VariableRefrigerantFlow to be connected to AirloopHVAC and AirLoopHVAC:OutdoorAirSystem:EquipmentList #7605

Merged

40 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Correct physics based VRF model to align inlet and outlet air flow rates #7295

Correct physics based VRF model to align inlet and outlet air flow rates #7295

rraustad commented May 9, 2019 •

edited

rraustad May 9, 2019

rraustad May 14, 2019

rraustad May 14, 2019

rraustad May 9, 2019

rraustad May 9, 2019 •

edited

rraustad May 9, 2019 •

edited

rraustad May 9, 2019

rraustad commented May 14, 2019 •

edited

rraustad May 14, 2019

rraustad May 14, 2019

rraustad May 14, 2019

rraustad May 14, 2019

rraustad May 15, 2019

Myoldmopar May 17, 2019

rraustad commented May 15, 2019 •

edited

rraustad May 15, 2019

Myoldmopar May 17, 2019

rraustad commented May 16, 2019

Myoldmopar commented May 17, 2019

Myoldmopar May 17, 2019

Myoldmopar May 17, 2019

Myoldmopar May 17, 2019

Myoldmopar May 17, 2019

Myoldmopar May 17, 2019

Myoldmopar May 17, 2019

Myoldmopar commented May 17, 2019

rraustad commented May 17, 2019

rraustad commented May 17, 2019 •

edited

Myoldmopar commented May 17, 2019

Myoldmopar commented May 17, 2019

rraustad commented May 17, 2019

Myoldmopar commented May 17, 2019

rraustad commented May 17, 2019

Myoldmopar commented May 17, 2019

Correct physics based VRF model to align inlet and outlet air flow rates #7295

Correct physics based VRF model to align inlet and outlet air flow rates #7295

Conversation

rraustad commented May 9, 2019 • edited

Pull request overview

Work Checklist

Review Checklist

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 May 9, 2019 • edited

Choose a reason for hiding this comment

rraustad May 9, 2019 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rraustad commented May 14, 2019 • edited

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 May 15, 2019 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rraustad commented May 16, 2019

Myoldmopar commented May 17, 2019

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 commented May 17, 2019

rraustad commented May 17, 2019

rraustad commented May 17, 2019 • edited

Myoldmopar commented May 17, 2019

Myoldmopar commented May 17, 2019

rraustad commented May 17, 2019

Myoldmopar commented May 17, 2019

rraustad commented May 17, 2019

Myoldmopar commented May 17, 2019

rraustad commented May 9, 2019 •

edited

rraustad May 9, 2019 •

edited

rraustad May 9, 2019 •

edited

rraustad commented May 14, 2019 •

edited

rraustad commented May 15, 2019 •

edited

rraustad commented May 17, 2019 •

edited