Are E+ToFMU 3.0.0 and 3.1.0 compatible backwards to older E+ versions?

[nicoca...@gmail.com]

Hi everyone, 

I used both versions of E+ToFMU 3.0.0 (python 2.7) and 3.1.0 (python (3.7) to create an FMU for an E+ model. I could successfully create the FMU and run the model in python without any problems. The main point was to upgrade my current FMU from CS1 to CS2. 

However, when I compare the results to the ones obtained by creating an FMU using E+ToFMU 2.1.0, the new ones are very strange. When I simulate the model using the EP Launcher, the results coincide with my CS1 model. I am using EnergyPlus 8.9.0, could this be the reason of the odd results? 

Thank you very much in advance. 

Best,

Nicolas.

[Thierry Nouidui]

Few questions:

• Is your E+ model from the EnergyPlusToFMU example folder? If yes, which example are you running?
• What do you mean by the results are "very strange"? 
• Could you post any figures comparing the old and new results?
• What master algorithm are you using to drive your FMU?

--
You received this message because you are subscribed to the Google Groups "energyplus-fmu" group.
To unsubscribe from this group and stop receiving emails from it, send an email to energyplus-fm...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/energyplus-fmu/a6703b9a-38a2-47f0-a412-41ee7b731b51n%40googlegroups.com.

--

Prof. Thierry Nouidui

Nouidui Consulting Group

https://nouidui.com/index.php/consulting/

[nicoca...@gmail.com]

Dear Thierry, 

thank you for your fast response. My answers: 

- No, it is a multizone building model (6 zones) that I use to run simulations. I used the same .idf file to create the FMU. 

- I am running the simulations in co-simulation mode using pyfmi (python 3.7). 

- Results very strange means that the don't match at all. The results should be the same, because the boundary conditions are the same (i am running the same script, just changing the loaded FMU file). The obtained results are logical for the FMU 1, not FMU 2. I attach here six pictures regarding room temperature, CO2 concentration and ventilation volume flow for both simulations. The internal loads are scheduled. The heating system is run by a PID using set points. The ventilation system controls the volume flow as a function of the indoor relative humidity and CO2. The obtained volume flows in the FMU 2 are very low (as a consequence of strangely low CO2 values). Consequently, the general air exchange rate is lower in the FMU 2, but the CO2 is also lower. Besides, the room temperatures are way below the set points, and the heating system is not able to keep them. I thought of the infiltration (I am using airflow network), but once again, the modeled infiltration is the same for both FMUs (same idf file).  

As I said, I thought that the problem might be backwards compatibility of E+ToFMU 3.x.x to earlier versions of E+ (8.x.x). The results obtained by the FMU 2 are the same if the FMU is created with python 2.7 (E+ToFMU 3.0.0) or 3.7 (E+ToFMU 3.1.0). I just wanted to upgrade to FMI 2.0 standard, but I can carry on simulating with the older one.  

Thank you very much for you help! 

Best,

Nicolas. 

[Thierry Nouidui]

Your results are indeed very puzzling.

What model/tool are you co-simulating your EnergyPlus FMU with?

Could you send the script (sections where the FMUs are loaded and run) that you are using to run the co-simulation for both cases (when you use FMU1 and FMU 2)?

Thanks!

To view this discussion on the web visit https://groups.google.com/d/msgid/energyplus-fmu/f0c76bb7-ffbb-446a-9792-1923a9d1c396n%40googlegroups.com.

[nicoca...@gmail.com]

Dear Thierry,

- I usually simulate my HVAC system in Modelica (using Dymola solvers). However, for this trial, I just run E+. I have some auxiliary functions in python to replace the missing HVAC model. 

- I attach you here the script. You will see a B everywhere (means one of the rooms, I simplified the script to simulate just a single room, but his repeats itself for every room). 

Thank you for the support! 

[Thierry Nouidui]

Thanks,

I am assuming that you are using the same script for running FMU 1.0 and FMU 2.0.

If that is the case, I wonder why for FMU 2.0 you don't call 

epfmu.setup_experiment(xxx) with the appropriate arguments to set-up the FMU (see the FMI standard or PyFMI for references) prior to calling the initialize method() as required by the FMI specification?

Please note that the start and stop time of the FMUs that will be run are obtained from that function.

It may be that you are simulating two different time periods depending on whether you are using FMU 1.0 and FMU 2.0.

This may justify the plots you are getting.

Thanks!

To view this discussion on the web visit https://groups.google.com/d/msgid/energyplus-fmu/2400eaa1-8bbe-441c-b44c-900ea367b5c6n%40googlegroups.com.

[nicoca...@gmail.com]

Hi Thierry,

I never called fmu.setup_experiment(xxx) because the documentation of the function initialize() says that it calls automatically the setup_experiment function. I had simulated several FMU 2.0 models before from Modelica, never called setup_experiment() but always obtained plausible results. 

I tried your recommendation but the results are the same. Maybe something else? 

Thanks!!! 

[Thierry Nouidui]

Could you please share the output folder of EnergyPlus FMU for both cases (1.0 and 2.0)? 

Also please let me know the simulation start time and end times?

Lastly, could you import those FMUs (1.0 and 2.0) in Dymola and run them using a different master to see if you can reproduce the problem?

You could set the inputs to some time dependent functions just to see how the outputs will be.

Thanks!

To view this discussion on the web visit https://groups.google.com/d/msgid/energyplus-fmu/da976fe8-d07d-4196-9ee1-d522bfb17c82n%40googlegroups.com.

[nicoca...@gmail.com]

The output folders weight around 100 MB each (with good compression a single file of 82 MB). Should I take out the results or do you want the full file?

[Thierry Nouidui]

Please take out the results first.

To view this discussion on the web visit https://groups.google.com/d/msgid/energyplus-fmu/49d0e67b-ee9b-424b-b459-025f4bd1f3cbn%40googlegroups.com.

[Thierry Nouidui]

Is the output from the script you have sent me?

Best will be to use the script you have sent to drive the FMU 1.0 and FMU 2.0 to maybe about a week of simulation assuming that the inputs will cause some changes to the outputs. 

Thank you

On Wed, Feb 17, 2021, 11:53 AM nicoca...@gmail.com <nicoca...@gmail.com> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/energyplus-fmu/49d0e67b-ee9b-424b-b459-025f4bd1f3cbn%40googlegroups.com.

[nicoca...@gmail.com]

Dear Thierry, 

I am sorry it took me so long to reply, I am full with other work stuff. Here you have the output files for the same simulation, using the same script and boundary conditions, using two different FMUs (1.0 and 2.0). If you need anything else please let me know. 

To download (Google is deleting my message when i attach the file): https://we.tl/t-TXmYJcsstw

Best,

Nicolas.

[Thierry Nouidui]

I am assuming that those two folders contain the results you have used to compare your two runs.

Looking at the EnergyPlus error message file in folder fmu_2_310\Output_EPExport_Model, I see that something went wrong with that simulation at time 168

 ** Severe  ** ExternalInterface: Socket communication received error value "-1" at time = 168.00 hours.
   **   ~~~   ** ExternalInterface: Flag from server " 0".
   **  Fatal  ** Error in ExternalInterface: Check EnergyPlus *.err file.
   ...Summary of Errors that led to program termination:
   ..... Reference severe error count=1
   .

This indicates that your simulation was not successful and hence the wrong results.

Generally speaking, you should always look at the error file first to know whether there was an issue with your run. 

You will have to debug and see why your simulation has failed at that particular time.

--
You received this message because you are subscribed to the Google Groups "energyplus-fmu" group.
To unsubscribe from this group and stop receiving emails from it, send an email to energyplus-fm...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/energyplus-fmu/fdc9cf8f-f06a-4b21-9d50-317ffd63eb0fn%40googlegroups.com.

[nicoca...@gmail.com]

I checked the error file before, and this is caused from an array that has the same length of the simulation steps, but somehow caused an error in the last simulation step (168 hs = 7 days, I simulated one week).

I had this error before also when running simulations with FMU 1.0, and did not alter the obtained simulations before this error. Do you think that this can change everything simulated until the error? This is why I ignored it.  

[Thierry Nouidui]

OK, If you are simulating for 7 days, then the problem is not related to the error I think.

I have looked at your file furthermore:

I made a diff of LXB_Build_FMU.expidf (in both Output folder of your runs) and see some difference in the following parameters/settings for the Windows

In the FMU 1.0 output folder (Output_EPExport_Slave) I see 

AirflowNetwork:MultiZone:Component:DetailedOpening,
Window Large,
0.0001,

AirflowNetwork:MultiZone:Component:DetailedOpening,
Window Medium,
0.0001,

In the FMU 2.0 output folder (Output_EPExport_Model) I see

AirflowNetwork:MultiZone:Component:DetailedOpening,
Window Large,
0.025,

AirflowNetwork:MultiZone:Component:DetailedOpening,
Window Medium,
0.025,

Why do you have such a difference in both models?

This difference is probably also reflected in your original IDF files since the expidf is from the original IDF.

To view this discussion on the web visit https://groups.google.com/d/msgid/energyplus-fmu/161b4e6e-2fb1-4bac-9b32-334a13b4f2edn%40googlegroups.com.