Failed to simulate moisture transfer in Spawn

[joyous-0130]

Hi Spawn and MBL developers,

I am trying to run the Buildings.ThermalZones.EnergyPlus_9_6_0.Validation.RefBldgSmallOffice with my own .idf file. In that .idf file, I just duplicated and change the HeatBalanceAlgorithm from 'ConductionTransferFunction' to 'MoisturePenetrationDepthConductionTranferFunction' and set the material property of the envelope that is necessary for mositure transfer. The .idf can simulate successfully in EnergyPlus V9.6.0. However, when simulate in the Modelica, it failed. I am running Dymola v2023 with Modelica Buildings Library v11.0.0. I am using Windows 11.

My question is: whether heat balance algorithm other than 'ConductionTransferFunction' are supported in Spwan.
Thank you so much. I would appreciate any insight.

Log-file of program ./dymosim
(generated: Tue Sep 17 20:21:53 2024)

dymosim started
... "Buildings.ThermalZones.EnergyPlus_9_6_0.Validation.ThermalZone.RefBldgSmallOffice" simulating
... "dsin.txt" loading (dymosim input file)
... loading "EnergyPlus" from "D:/Modelica/Buildings 11.0.0/Resources/Data/ThermalZones/EnergyPlus_9_6_0/Examples/RefBldgSmallOffice/RefBldgSmallOfficeNew2004_Chicago.dat"
0.000 RefBldgSmallOffice.building: Error from EnergyPlus: Convergence error in SolveForWindowTemperatures for window PERIMETER_ZN_1_WALL_SOUTH_WINDOW_1
0.000 RefBldgSmallOffice.building: Error from EnergyPlus:  During Warmup, Environment=SPAWN-RUNPERIOD, at Simulation time=01/01 00:18 - 00:18
0.000 RefBldgSmallOffice.building: Fatal from EnergyPlus: Program halted because of convergence error in SolveForWindowTemperatures for window PERIMETER_ZN_1_WALL_SOUTH_WINDOW_1
0.000 RefBldgSmallOffice.building: Fatal from EnergyPlus: EnergyPlus is not running
Error: The following error was detected at time: 0
Failed to get parameter values for RefBldgSmallOffice.wes. This may be due to an error during the initialization or warm-up of EnergyPlus as the EnergyPlus FMU has been generated and loaded with no error.

The stack of functions is:
Buildings.ThermalZones.EnergyPlus_9_6_0.BaseClasses.initialize
Buildings.ThermalZones.EnergyPlus_9_6_0.BaseClasses.initialize(
wes.fmuZon.adapter, 
building.isSynchronized)

Error: Integrator failed to start model.

... Error message from dymosim

ERROR: The simulation of Buildings.ThermalZones.EnergyPlus_9_6_0.Validation.ThermalZone.RefBldgSmallOffice FAILED

[mwetter]

@kbenne : Can you please confirm that MoisturePenetrationDepthConductionTranferFunction is supported, and do you have any insight into what triggers this error in SolveForWindowTemperatures.

[kbenne]

We aren't explicitly testing this scenario so I can't be completely confident that it will bahave as expected. I see that the problem appears to arise during warmup, therefore one thing that comes to mind is that Spawn uses additional warmup convergence criteria than vanilla EnergyPlus. In Spawn, the convergence criteria looks for periodic stability of the surface temperatures in addition to stabliity of the air temperature. In contrast, EnergyPlus only checks for stability of the air temperature during warmup. This means that Spawn will likely iterate longer during warmup.

There are further complications if there are EnergyPlus surfaces that are actuated from the Modelica side. I'm assuming this is not the case, but it is a scenario that warrants some consideration, since there is nothing preventing someone from creating a model with actuated surfaces and MoisturePenetrationDepthConductionTranferFunction model.

As a first step in troubleshooting, I would try to run the model in Modelica, but without any EnergyPlus zones connected to Modelica, or Modelica inputs of any sort. The zones of the EnergyPlus model will simply float according to the load and ambient conditions in this configuration. Admittately, this configuation iteself, may create differences in warmup between EnergyPlus and Spawn, but I think it is a good test.