Within the Modelica model one can set some default values for parameters. Also, one can define variables for output. However, for a very complicated model, sometimes it is very difficult to output every desired variable, e.g., protected variables. Therefore, is it possible to modify parameter values and add output variables using Python code or the fmpy.gui?

I'm generating a Co-Simulation FMU using Dymola. The FMU is consumed by FMPy. There are external thermodynamic calls in the model.

When the Dymola flag Advanced.ParallelizeCode is set to false, the FMU can be run using FMPy and Dymola. When the Dymola flag Advanced.ParallelizeCode is set to true, the FMU can be run using Dymola but not using FMPy.

What can be causing this issue?

Thanks!

The "Unknown" elements in "ModelStructure/Outputs" and "ModelStructure/Derivatives" are currently not read from the modelDescription.xml.

In FMPy 0.2.3., writing new parameter values to the xml file has been introduced.
However, this is dangerous, as it is not specified, whether an importing tool should consider these values or the hard coded values, see this discussion: [https://trac.fmi-standard.org/ticket/387].
Strictly speaking, changing the modeldescription.xml could even invalidate the FMU (consistency of xml and binary could be checked by creating a GUID with by a checksum.
The clean way to support changed parameter and even different parameter sets is the coming SSP standard.

Looking at the source code for command_line.py, (specifically this line), I don't see how this command line parameter has any effect. It doesn't appear as though the value extracted from the command line is used anywhere.

Am I missing something?

Hi,

I am facing an issue (dll not unlinked, see error message below) with FMUs generated with Dymola (tested with Versions 2017FD01, 2018FD01). The simple example below works fine, when using the FMU provided by the following link:

https://trac.fmi-standard.org/browser/branches/public/Test_FMUs/FMI_2.0/ModelExchange/win64/Dymola/2017/CoupledClutches/CoupledClutches.fmu

I am working with Python 3.6.2 (see installed packages below) and I am using the FMPy develop branch (the problem also occurs on the master branch release 0.2.1).

It seems that the dll is not unlinked properly. Inspecting threads I found the following line:

Modelica_Mechanics_Rotational_Examples_CoupledClutches_DYMOLA2017FD01.dll!fmi2GetStringStatus+0x5bbec

Thanks!
Alexander

Export FMU (checked settings):

Type: Model Exchange
Version: 2.0
Binaries: 32-bit, 64-bit
Model description filters: Protected variables
Options: Copy resources to FMU, Show dialog when exporting

Simple example:

from fmpy import simulate_fmu

#filename = 'fmus/Modelica_Mechanics_Rotational_Examples_CoupledClutches_DYMOLA2018FD01.fmu'
filename = 'fmus/Modelica_Mechanics_Rotational_Examples_CoupledClutches_DYMOLA2017FD01.fmu'
# filename = 'fmus/Dymola2017/CoupledClutches.fmu'

simulate_fmu(
    filename=filename,
    start_time=0,
    stop_time=1.5,
    solver='CVode',
    step_size=1e-2,
    output_interval=2e-2,
    input=None,
    output=None,
    validate=False,
    fmi_logging=False)

Error message:

---------------------------------------------------------------------------
PermissionError                           Traceback (most recent call last)
<ipython-input-11-a27bbb9b9121> in <module>()
     17     output=None,
     18     validate=False,
---> 19     fmi_logging=False)

c:\users\opa2si1\documents\git_projects\fmpy\fmpy\simulation.py in simulate_fmu(filename, validate, start_time, stop_time, solver, step_size, relative_tolerance, output_interval, fmi_type, use_source_code, start_values, input, output, timeout, fmi_logging, logger, step_finished)
    435 
    436     # clean up
--> 437     shutil.rmtree(unzipdir)
    438 
    439     return result

~\AppData\Local\conda\conda\envs\FMPy_issue_PermissionError\lib\shutil.py in rmtree(path, ignore_errors, onerror)
    492             os.close(fd)
    493     else:
--> 494         return _rmtree_unsafe(path, onerror)
    495 
    496 # Allow introspection of whether or not the hardening against symlink

~\AppData\Local\conda\conda\envs\FMPy_issue_PermissionError\lib\shutil.py in _rmtree_unsafe(path, onerror)
    382             mode = 0
    383         if stat.S_ISDIR(mode):
--> 384             _rmtree_unsafe(fullname, onerror)
    385         else:
    386             try:

~\AppData\Local\conda\conda\envs\FMPy_issue_PermissionError\lib\shutil.py in _rmtree_unsafe(path, onerror)
    382             mode = 0
    383         if stat.S_ISDIR(mode):
--> 384             _rmtree_unsafe(fullname, onerror)
    385         else:
    386             try:

~\AppData\Local\conda\conda\envs\FMPy_issue_PermissionError\lib\shutil.py in _rmtree_unsafe(path, onerror)
    387                 os.unlink(fullname)
    388             except OSError:
--> 389                 onerror(os.unlink, fullname, sys.exc_info())
    390     try:
    391         os.rmdir(path)

~\AppData\Local\conda\conda\envs\FMPy_issue_PermissionError\lib\shutil.py in _rmtree_unsafe(path, onerror)
    385         else:
    386             try:
--> 387                 os.unlink(fullname)
    388             except OSError:
    389                 onerror(os.unlink, fullname, sys.exc_info())

PermissionError: [WinError 5] Access is denied: 'C:\\Users\\xxxxxxx\\AppData\\Local\\Temp\\tmp4v696d0o\\binaries\\win64\\Modelica_Mechanics_Rotational_Examples_CoupledClutches_DYMOLA2017FD01.dll'

requirements (platform: win-64):

bleach=1.5.0=py36_0
ca-certificates=2018.1.18=0
certifi=2016.2.28=py36_0
colorama=0.3.9=py36_0
decorator=4.1.2=py36_0
entrypoints=0.2.3=py36_0
html5lib=0.9999999=py36_0
icu=58.2=vc14_0
ipykernel=4.6.1=py36_0
ipython=6.1.0=py36_0
ipython_genutils=0.2.0=py36_0
ipywidgets=6.0.0=py36_0
jedi=0.10.2=py36_2
jinja2=2.9.6=py36_0
jpeg=9b=vc14_2
jsonschema=2.6.0=py36_0
jupyter=1.0.0=py36_3
jupyter_client=5.1.0=py36_0
jupyter_console=5.2.0=py36_0
jupyter_core=4.3.0=py36_0
libpng=1.6.34=vc14_0
lxml=3.8.0=py36_0
markupsafe=1.0=py36_0
mistune=0.7.4=py36_0
mkl=2017.0.3=0
nbconvert=5.2.1=py36_0
nbformat=4.4.0=py36_0
notebook=5.0.0=py36_0
numpy=1.13.1=py36_0
openssl=1.0.2n=vc14_0
pandocfilters=1.4.2=py36_0
path.py=10.3.1=py36_0
pathlib=1.0.1=py36_0
pickleshare=0.7.4=py36_0
pip=9.0.1=py36_1
prompt_toolkit=1.0.15=py36_0
pygments=2.2.0=py36_0
pyqt=5.6.0=py36_2
python=3.6.2=0
python-dateutil=2.6.1=py36_0
pywin32=220=py36_2
pyzmq=16.0.2=py36_0
qt=5.6.2=vc14_1
qtconsole=4.3.1=py36_0
setuptools=36.4.0=py36_1
simplegeneric=0.8.1=py36_1
sip=4.18=py36_0
six=1.10.0=py36_0
testpath=0.3.1=py36_0
tornado=4.5.2=py36_0
traitlets=4.3.2=py36_0
vc=14=0
vs2015_runtime=14.0.25420=0
wcwidth=0.1.7=py36_0
wheel=0.29.0=py36_0
widgetsnbextension=3.0.2=py36_0
wincertstore=0.2=py36_0
zlib=1.2.11=vc14_0

Can't use fmpy.gui under Python 2.7.12 due to:

Traceback (most recent call last):
  File "/home/user/.local/lib/python2.7/site-packages/fmpy/gui/__main__.py", line 468, in open
    self.load(filename)
  File "/home/user/.local/lib/python2.7/site-packages/fmpy/gui/__main__.py", line 399, in load
    self.treeModel.setModelDescription(md)
  File "/home/user/.local/lib/python2.7/site-packages/fmpy/gui/model.py", line 197, in setModelDescription
    self.rootItem.children.clear()
AttributeError: 'list' object has no attribute 'clear'
Aborted (core dumped)

If the user specifies an invalid fmi_type, e.g. fmi_type='CoSimulation for a ME-only FMU, this is not caught during the argument checks below:

Sorry, I can't provide a PR yet, but the fix is easy by checking if the FMU really supports the given mode, e.g.:

if fmi_type is not None:
    if (fmi_type == 'CoSimulation' and not modelDescription.coSimulation) or \
            (fmi_type == 'ModelExchange' and not modelDescription.modelExchange):
        raise ValueError("Invalid choice. FMU doesn't support fmi_type=%r" % fmi_type)

There is a comment in the model_description.py file that notes:
#TODO: add enumeration items

While at first glance this appears to be a relatively insignificant feature, it would absolutely be a benefit to users if it could be added.

As an example, I am working on a tool to run FMUs within Python using FMPy, and then present the results to a user who may not be familiar with the code/architecture of the FMU. To provide clear results to the user, it is important to be able to display enumeration values not only as their integer values but also as their 'name' or 'description' strings so that the user gains insight.

I am still just learning Python or I would try to tackle this problem myself, but I would appreciate if you have the time to work at this feature.

Thank you,
Justin

Is there a possibility to install the package without needing an interent connection? I have tried all the possible ways to install (apart from PIP) and all have failed due to connect timeouts. I suppose the problem in my case lies in the firewall of the company or in a proxy setting.
What is an aternative to this? Can one download the package and just install it?
Thanks in advance for any hint.

Emmanuel

PyPI doesn't seem to support rendering markdown, while rst would work both for Github and PyPi.

http://stackoverflow.com/questions/26737222/pypi-description-markdown-doesnt-work

Hi,
I'm using FMPy for my research. Is there a specific format to cite this tool? For example in BibTex format?

Hey
i realized that FMPy doesn't bring the option to serialize an FMU state. It would be really nice if you implement this feature FMI Feature (fmi2SerializeFMUstate) because the FMU states couldn't be stored by pickle. Or do you have any other idea how to store them?

Hi, in the coupled_clutches.py example one can set the start value using the simulate_fmu functionality. Is there a similar way to modify the start value (of a parameter) for an FMU instance? For example, how to modify the start value for the FMU in the custom_input.py example? To be specific, the parameter might have the following property, and I would like to modify the value 1.
<ScalarVariable name="friction" valueReference="111" description="Friction coefficient" causality="parameter" variability="tunable"> <Real start="1"/>

https://trac.fmi-standard.org does not contain Rectifier.fmu for linux64/Dymola.
When I try to simulate parameter_variation.py on Linux I get the following error:

Exception: Failed to download https://trac.fmi-standard.org/export/HEAD/branches/public/Test_FMUs/FMI_2.0/CoSimulation/linux64/Dymola/2017/Rectifier/Rectifier.fmu (status code: 404)

Other examples run fine. It would be helpful to mark OS-dependent tests and maybe disable them on unsupported platforms.

(Thanks for your work on this package 👍 )

Is there any way to update the parameter value (such as k, Ti, Td) within the PID block, in the Simulate SPPs example?

I notice that simulating an FMU with no outputs creates a 'Bad file descriptor' exception in Ubuntu 16.04, and it leaves the interpreter hanging.

To replicate the problem, this is the FMU file (renamed to .fmu.zip to comply to GitHub's restrictions on attaching files to issues) NoOutputSystem.fmu.zip that is creating the problem, that was generated using OpenModelica from this Modelica model:

model NoOutputSystem
  Real x "An unknown";
initial equation
  x = 0.9 "Used before simulation to compute initial values";
equation
  der(x) = -x+x*x*x;
  annotation(experiment(StartTime=0,StopTime=10));
end NoOutputSystem;

This is the error:

traversaro@turati:~/: ipython
Python 2.7.12 (default, Nov 20 2017, 18:23:56) 
In [1]: import fmpy

In [2]: fmu = "NoOutputSystem.fmu"

In [3]: fmpy.simulate_fmu(fmu)
Out[3]: 
array([(0.02,), (0.04,), (0.08,), (0.1,), (0.12000000000000001,), (0.14,),
       (0.18,), (0.21999999999999997,), (0.25999999999999995,),
       (0.27999999999999997,), (0.3,), (0.32,), (0.34,),
       (0.36000000000000004,), (0.38000000000000006,),
       (0.4000000000000001,), (0.4200000000000001,), (0.4400000000000001,),
       (0.46000000000000013,), (0.48000000000000015,),
       (0.5200000000000001,), (0.5400000000000001,), (0.5600000000000002,),
       (0.5800000000000002,), (0.6000000000000002,), (0.6200000000000002,),
       (0.6400000000000002,), (0.6600000000000003,), (0.6800000000000003,),
       (0.7000000000000003,), (0.7200000000000003,), (0.7400000000000003,),
       (0.7600000000000003,), (0.7800000000000004,), (0.8000000000000004,),
       (0.8200000000000004,), (0.8400000000000004,), (0.8600000000000004,),
       (0.8800000000000004,), (0.9000000000000005,), (0.9200000000000005,),
       (0.9400000000000005,), (0.9600000000000005,), (0.9800000000000005,),
       (1.0200000000000005,), (1.0400000000000005,), (1.0600000000000005,),
       (1.0800000000000005,), (1.1000000000000005,), (1.1200000000000006,),
       (1.1400000000000006,), (1.1600000000000006,), (1.1800000000000006,),
       (1.2000000000000006,), (1.2200000000000006,), (1.2400000000000007,),
       (1.2600000000000007,), (1.2800000000000007,), (1.3000000000000007,),
       (1.3200000000000007,), (1.3400000000000007,), (1.3600000000000008,),
       (1.3800000000000008,), (1.4000000000000008,), (1.4200000000000008,),
       (1.4400000000000008,), (1.4600000000000009,), (1.4800000000000009,),
       (1.5000000000000009,), (1.520000000000001,), (1.540000000000001,),
       (1.560000000000001,), (1.580000000000001,), (1.600000000000001,),
       (1.620000000000001,), (1.640000000000001,), (1.660000000000001,),
       (1.680000000000001,), (1.700000000000001,), (1.720000000000001,),
       (1.740000000000001,), (1.7600000000000011,), (1.7800000000000011,),
       (1.8000000000000012,), (1.8200000000000012,), (1.8400000000000012,),
       (1.8600000000000012,), (1.8800000000000012,), (1.9000000000000012,),
       (1.9200000000000013,), (1.9400000000000013,), (1.9600000000000013,),
       (1.9800000000000013,), (2.0000000000000013,), (2.0200000000000014,),
       (2.0400000000000014,), (2.0600000000000014,), (2.0800000000000014,),
       (2.1000000000000014,), (2.1200000000000014,), (2.1400000000000015,),
       (2.1600000000000015,), (2.1800000000000015,), (2.2000000000000015,),
       (2.2200000000000015,), (2.2400000000000015,), (2.2600000000000016,),
       (2.2800000000000016,), (2.3000000000000016,), (2.3200000000000016,),
       (2.3400000000000016,), (2.3600000000000017,), (2.3800000000000017,),
       (2.4000000000000017,), (2.4200000000000017,), (2.4400000000000017,),
       (2.4600000000000017,), (2.4800000000000018,), (2.5000000000000018,),
       (2.520000000000002,), (2.540000000000002,), (2.560000000000002,),
       (2.580000000000002,), (2.600000000000002,), (2.620000000000002,),
       (2.640000000000002,), (2.660000000000002,), (2.680000000000002,),
       (2.700000000000002,), (2.720000000000002,), (2.740000000000002,),
       (2.760000000000002,), (2.780000000000002,), (2.800000000000002,),
       (2.820000000000002,), (2.840000000000002,), (2.860000000000002,),
       (2.880000000000002,), (2.900000000000002,), (2.920000000000002,),
       (2.940000000000002,), (2.960000000000002,), (2.980000000000002,),
       (3.000000000000002,), (3.0200000000000022,), (3.0400000000000023,),
       (3.0600000000000023,), (3.0800000000000023,), (3.1000000000000023,),
       (3.1200000000000023,), (3.1400000000000023,), (3.1600000000000024,),
       (3.1800000000000024,), (3.2000000000000024,), (3.2200000000000024,),
       (3.2400000000000024,), (3.2600000000000025,), (3.2800000000000025,),
       (3.3000000000000025,), (3.3200000000000025,), (3.3400000000000025,),
       (3.3600000000000025,), (3.3800000000000026,), (3.4000000000000026,),
       (3.4200000000000026,), (3.4400000000000026,), (3.4600000000000026,),
       (3.4800000000000026,), (3.5000000000000027,), (3.5200000000000027,),
       (3.5400000000000027,), (3.5600000000000027,), (3.5800000000000027,),
       (3.6000000000000028,), (3.6200000000000028,), (3.640000000000003,),
       (3.660000000000003,), (3.680000000000003,), (3.700000000000003,),
       (3.720000000000003,), (3.740000000000003,), (3.760000000000003,),
       (3.780000000000003,), (3.800000000000003,), (3.820000000000003,),
       (3.840000000000003,), (3.860000000000003,), (3.880000000000003,),
       (3.900000000000003,), (3.920000000000003,), (3.940000000000003,),
       (3.960000000000003,), (3.980000000000003,), (4.020000000000002,),
       (4.060000000000001,), (4.1000000000000005,), (4.14,),
       (4.179999999999999,), (4.219999999999998,), (4.259999999999997,),
       (4.299999999999996,), (4.339999999999995,), (4.379999999999995,),
       (4.419999999999994,), (4.459999999999993,), (4.499999999999992,),
       (4.539999999999991,), (4.57999999999999,), (4.6199999999999894,),
       (4.659999999999989,), (4.699999999999988,), (4.739999999999987,),
       (4.779999999999986,), (4.819999999999985,), (4.859999999999984,),
       (4.8999999999999835,), (4.939999999999983,), (4.979999999999982,),
       (5.019999999999981,), (5.05999999999998,), (5.099999999999979,),
       (5.139999999999978,), (5.1799999999999775,), (5.219999999999977,),
       (5.259999999999976,), (5.299999999999975,), (5.339999999999974,),
       (5.379999999999973,), (5.419999999999972,), (5.4599999999999715,),
       (5.499999999999971,), (5.53999999999997,), (5.579999999999969,),
       (5.619999999999968,), (5.659999999999967,), (5.699999999999966,),
       (5.739999999999966,), (5.779999999999965,), (5.819999999999964,),
       (5.859999999999963,), (5.899999999999962,), (5.939999999999961,),
       (5.9799999999999605,), (6.01999999999996,), (6.059999999999959,),
       (6.099999999999958,), (6.139999999999957,), (6.179999999999956,),
       (6.219999999999955,), (6.2599999999999545,), (6.299999999999954,),
       (6.339999999999953,), (6.379999999999952,), (6.419999999999951,),
       (6.45999999999995,), (6.499999999999949,), (6.5399999999999485,),
       (6.579999999999948,), (6.619999999999947,), (6.659999999999946,),
       (6.699999999999945,), (6.739999999999944,), (6.779999999999943,),
       (6.8199999999999426,), (6.859999999999942,), (6.899999999999941,),
       (6.93999999999994,), (6.979999999999939,), (7.019999999999938,),
       (7.059999999999937,), (7.099999999999937,), (7.139999999999936,),
       (7.179999999999935,), (7.219999999999934,), (7.259999999999933,),
       (7.299999999999932,), (7.3399999999999315,), (7.379999999999931,),
       (7.41999999999993,), (7.459999999999929,), (7.499999999999928,),
       (7.539999999999927,), (7.579999999999926,), (7.6199999999999255,),
       (7.659999999999925,), (7.699999999999924,), (7.739999999999923,),
       (7.779999999999922,), (7.819999999999921,), (7.85999999999992,),
       (7.8999999999999195,), (7.939999999999919,), (7.979999999999918,),
       (8.019999999999918,), (8.059999999999917,), (8.099999999999916,),
       (8.139999999999915,), (8.179999999999914,), (8.219999999999914,),
       (8.259999999999913,), (8.299999999999912,), (8.339999999999911,),
       (8.37999999999991,), (8.41999999999991,), (8.459999999999908,),
       (8.499999999999908,), (8.539999999999907,), (8.579999999999906,),
       (8.619999999999905,), (8.659999999999904,), (8.699999999999903,),
       (8.739999999999903,), (8.779999999999902,), (8.8199999999999,),
       (8.8599999999999,), (8.899999999999899,), (8.939999999999898,),
       (8.979999999999897,), (9.019999999999897,), (9.059999999999896,),
       (9.099999999999895,), (9.139999999999894,), (9.179999999999893,),
       (9.219999999999892,), (9.259999999999891,), (9.29999999999989,),
       (9.33999999999989,), (9.379999999999889,), (9.419999999999888,),
       (9.459999999999887,), (9.499999999999886,), (9.539999999999885,),
       (9.579999999999885,), (9.619999999999884,), (9.659999999999883,),
       (9.699999999999882,), (9.739999999999881,), (9.77999999999988,),
       (9.81999999999988,), (9.859999999999879,), (9.899999999999878,),
       (9.939999999999877,), (9.979999999999876,), (10.0,), (10.0,)], 
      dtype=[('time', '<f8')])

In [4]: Exception in thread Thread-38:
Traceback (most recent call last):
  File "/usr/lib/python2.7/threading.py", line 801, in __bootstrap_inner
    self.run()
  File "/usr/lib/python2.7/threading.py", line 754, in run
    self.__target(*self.__args, **self.__kwargs)
  File "/usr/local/lib/python2.7/dist-packages/prompt_toolkit/eventloop/inputhook.py", line 68, in thread
    input_is_ready_func(wait=True)
  File "/usr/local/lib/python2.7/dist-packages/prompt_toolkit/eventloop/posix.py", line 101, in ready
    return self._ready_for_reading(current_timeout[0] if wait else 0) != []
  File "/usr/local/lib/python2.7/dist-packages/prompt_toolkit/eventloop/posix.py", line 188, in _ready_for_reading
    fds = self.selector.select(timeout)
  File "/usr/local/lib/python2.7/dist-packages/prompt_toolkit/eventloop/select.py", line 100, in select
    return self._select_selector.select(timeout)
  File "/usr/local/lib/python2.7/dist-packages/prompt_toolkit/eventloop/select.py", line 182, in select
    return select.select(self._fds, [], [], timeout)[0]
error: (9, 'Bad file descriptor')

^[[70;1R

The same example works fine by turning "x" in a output variable.

Everything in the title. It seems important for me to be able to run 32 bits fmu on windows 64bits (most common OS)

Although the project can be found on https://testpypi.python.org/pypi/fmpy/0.0.1

installation per

pip install fmpy

still doesn't work for me (on Linux with Python 3.6)

I see that one can simulate several FMUs with SSP. Is there a similar procedure for closed loop simulation? For example, a controller designed directly in Python and an FMU plant?

Hi,

When trying to open the gui, it gives the following error.
No module named fmpy.gui.

Also, in the installation folder, there is no gui folder as the one in github.

Any idea why?

Thanks in advance.

While exploring the examples a ran into the following issue:

python -m fmpy.examples.coupled_clutches

Hangs at:

  Downloading https://trac.fmi-standard.org/export/HEAD/branches/public/Test_FMUs/FMI_2.0/ModelExchange/darwin64/Maple
  Sim/2016.2/CoupledClutches/CoupledClutches.fmu
  Downloading https://trac.fmi-standard.org/export/HEAD/branches/public/Test_FMUs/FMI_2.0/ModelExchange/darwin64/Maple
  Sim/2016.2/CoupledClutches/CoupledClutches_in.csv
  Loading input...
  Simulating CoupledClutches.fmu (FMI 2.0, ModelExchange, CVode)...
  Plotting results...

I added an extra import statement and matplotlib.use() call like suggested here which solved the issue.

...
import matplotlib
matplotlib.use('TkAgg')
...

Running Mac OS 10.13.2 (17C205)

The simulate_fmu function has a named input argument. However, it isn't clear what type of data is required there. The in-source documentation simple says a structured numpy array that contains the input, but that doesn't mean much to me.

I eventually figured out by tracing back through the Input class and the np.genfromtxt functions. But it would be nice to have something more explicit for how to build the data that is required.

Hi,

While trying to run the example with the option "Simulate an FMU in Python", it gives the following error.

plot_result(result)
Traceback (most recent call last):
File "", line 1, in
File "C:\Python27\lib\site-packages\fmpy\util.py", line 299, in plot_result
fig.set_size_inches(w=8, h=1.5 * len(names), forward=True)
File "C:\Python27\lib\site-packages\matplotlib\figure.py", line 667, in set_size_inches
w, h = args
ValueError: need more than 0 values to unpack

Hi,

Though I am looking through the examples and source codes, it'd be great to read through actual API documentation, but I can't find one.

Thanks!

It would be beneficial to be able to set the absolute and relative tolerances for the CVode Solver and not use default values.

Hi Torsten,

in the current implementation, during every call of 'simulate_fmu()', the fmu is unzipped, intialized and after the run terminated and the temporary folder is deleted.
This is problematic for several (thousand) runs with the same fmu, e.g. Monte-Carlo. Would it be possible to re-structure and split the simulation in a init-, a run- and a terminate call, e.g. as a class:

class simulate_with_fmu(object):

   def __init__(self):
     unzipdir = extract(filename)
     fmu.initialize()
     ...

   def run():
      return simulateME()

   def __exit__(self):
       fmu.terminate()
       fmu.freeInstance()
       shutil.rmtree(unzipdir)

Hi,

I'm trying to control the simulation progress with Python, so I generated a fmu file of StableBouncingBall and import with FMPy. When I run simulate_fmu, I got an error "The current platform (win32) is not supported by the FMU". Then I read some lines of source file and added "use_source_code=True", hoping that FMPy will simulate with source-code. Then I got another error "No source files specified in the model description." But I unziped the fmu file and found .c and .h source file in the model.

The fmu file works fine in Matlab 2018 for both Co-Simulation and Model Exchange. Is there any way to solve this problem?

Best regards,
XY Fan

The following code fails because all requests to https://trac.fmi-standard.org/ are now redirected to https://github.com/modelica/fmi-standard/issues/.

from fmpy.util import download_test_file

download_test_file('FMI_2.0', 'CoSimulation', 'Dymola', '2017', 'Rectifier', 'Rectifier.fmu')

This is also breaking some tests.

Hello,

I simulated ModelCosimulation on Raspberry Pi and it work's very well. Now I tried to simulate a ModelExchange on Raspberry and it doesn't work.

I used following file /FMI_2.0/ModelExchange/c-code/Dymola/2017/CoupledClutches.fmu

Compilation worked with

fmpy compile CoupledClutches.fmu
By simulation I get some error

$ fmpy simulate CoupledClutches.fmu --show-plot
Traceback (most recent call last):
  File "/usr/local/bin/fmpy", line 11, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.5/dist-packages/fmpy/command_line.py", line 114, in main
    fmi_call_logger=fmi_call_logger)
  File "/usr/local/lib/python3.5/dist-packages/fmpy/simulation.py", line 507, in simulate_fmu
    result = simulateME(model_description, fmu_args, start_time, stop_time, solver, step_size, relative_tolerance, start_values, apply_default_start_values, input, output, output_interval, record_events, timeout, callbacks, debug_logging, step_finished)
  File "/usr/local/lib/python3.5/dist-packages/fmpy/simulation.py", line 595, in simulateME
    from .sundials import CVodeSolver
  File "/usr/local/lib/python3.5/dist-packages/fmpy/sundials/__init__.py", line 16, in <module>
    sundials_nvecserial = cdll.LoadLibrary(os.path.join(library_dir, 'sundials_nvecserial' + sharedLibraryExtension))
  File "/usr/lib/python3.5/ctypes/__init__.py", line 425, in LoadLibrary
    return self._dlltype(name)
  File "/usr/lib/python3.5/ctypes/__init__.py", line 347, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: /usr/local/lib/python3.5/dist-packages/fmpy/sundials/linux32/sundials_nvecserial.so: cannot open shared object file: No such file or directory

This is my directory:

pi@server:/usr/local/lib/python3.5/dist-packages/fmpy/sundials $ tree
.
├── darwin64
│   ├── sundials_cvode.dylib
│   └── sundials_nvecserial.dylib
├── __init__.py
├── linux64
│   ├── sundials_cvode.so
│   └── sundials_nvecserial.so
├── __pycache__
│   └── __init__.cpython-35.pyc
├── win32
│   ├── sundials_cvode.dll
│   └── sundials_nvecserial.dll
└── win64
    ├── sundials_cvode.dll
    └── sundials_nvecserial.dll

I googled to sundials_cvode.so and sundials_nvecserial.so for Linux 32Bit but didn't find it.
How could I solve this problem?

Best regards

Hi,

I need the results to reside on a regular grid and I was expecting this behaviour since I explicitly specified the output_interval. Observing the actual output I found three issues, where I am not sure if this is intended by the author:

1. doesn't start at t=0s,
2. some samples are droped, e.g. at t=0.06s,
3. grid is shifted starting after t=0.42s.

Could you please clarify? Thanks.

Simple example:

from fmpy import simulate_fmu

filename = 'fmus/Dymola2017/CoupledClutches.fmu'

print(filename)
results = simulate_fmu(
    filename=filename,
    start_time=0,
    stop_time=1.5,
    solver='CVode',
    step_size=1e-2,
    output_interval=2e-2,
    input=None,
    output=None,
    validate=False,
    fmi_logging=False)

print(results['time'])

Output:

[ 0.02        0.04        0.08        0.1         0.12        0.14        0.18
  0.22        0.26        0.28        0.3         0.32        0.34        0.36
  0.38        0.4         0.42        0.44565285  0.46565285  0.49740768
  0.53740768  0.55740768  0.57740768  0.59740768  0.63429745  0.65429745
  0.67429745  0.70976556  0.72976556  0.74976556  0.76976556  0.78976556
  0.82000019  0.85116724  0.87116724  0.89619621  0.92065559  0.94296037
  0.96296037  0.98296037  1.00355202  1.03929382  1.05929382  1.07929382
  1.11029498  1.13029498  1.15733956  1.17733956  1.21676477  1.23676477
  1.27071325  1.29071325  1.31071325  1.33071325  1.3697638   1.3897638
  1.4097638   1.4297638   1.4497638   1.4701323   1.4901323   1.5       ]

I'm currently trying to simulate a MIMO systems with varying inputs and wrapped FMPy inside a simple class.

However, if I assign more than one input value to a model exchange simulation via a structured numpy array, I get the following error message:

ValueError                                Traceback (most recent call last)
<ipython-input-11-6f1be21121d8> in <module>()
----> 1 simulation.simulate_exchange(use_inputs=True)

/media/jmartensen/Data/general/test/fmu_test/FMPY_Wrapper.py in simulate_exchange(self, plot, use_inputs)
     75             input = self.input_array if use_inputs else None ,
     76             output = self.outputs,
---> 77             fmi_type = 'ModelExchange'
     78         )
     79 

~/miniconda3/envs/python36/lib/python3.6/site-packages/fmpy/simulation.py in simulate_fmu(filename, validate, start_time, stop_time, solver, step_size, relative_tolerance, output_interval, record_events, fmi_type, use_source_code, start_values, apply_default_start_values, input, output, timeout, debug_logging, logger, fmi_call_logger, step_finished, model_description)
    474     if fmi_type == 'ModelExchange' and model_description.modelExchange is not None:
    475         fmu_args['modelIdentifier'] = model_description.modelExchange.modelIdentifier
--> 476         result = simulateME(model_description, fmu_args, start_time, stop_time, solver, step_size, relative_tolerance, start_values, apply_default_start_values, input, output, output_interval, record_events, timeout, callbacks, debug_logging, step_finished)
    477     elif fmi_type == 'CoSimulation' and model_description.coSimulation is not None:
    478         fmu_args['modelIdentifier'] = model_description.coSimulation.modelIdentifier

~/miniconda3/envs/python36/lib/python3.6/site-packages/fmpy/simulation.py in simulateME(model_description, fmu_kwargs, start_time, stop_time, solver_name, step_size, relative_tolerance, start_values, apply_default_start_values, input_signals, output, output_interval, record_events, timeout, callbacks, debug_logging, step_finished)
    529 
    530     input = Input(fmu, model_description, input_signals)
--> 531     input.apply(time)
    532 
    533     if is_fmi1:

~/miniconda3/envs/python36/lib/python3.6/site-packages/fmpy/simulation.py in apply(self, time, continuous, discrete, after_event)
    192                              cast(self.boolean_values, POINTER(self._bool_type)))
    193 
--> 194         return Input.nextEvent(time, t)
    195 
    196     @staticmethod

~/miniconda3/envs/python36/lib/python3.6/site-packages/fmpy/simulation.py in nextEvent(time, t)
    201 
    202 
--> 203         i = np.argmax(t_events > time)
    204 
    205         if t_events[i] > time:

~/miniconda3/envs/python36/lib/python3.6/site-packages/numpy/core/fromnumeric.py in argmax(a, axis, out)
   1002 
   1003     """
-> 1004     return _wrapfunc(a, 'argmax', axis=axis, out=out)
   1005 
   1006 

~/miniconda3/envs/python36/lib/python3.6/site-packages/numpy/core/fromnumeric.py in _wrapfunc(obj, method, *args, **kwds)
     50 def _wrapfunc(obj, method, *args, **kwds):
     51     try:
---> 52         return getattr(obj, method)(*args, **kwds)
     53 
     54     # An AttributeError occurs if the object does not have

ValueError: attempt to get argmax of an empty sequence

This happens as well if I try to simulate the fmu via fmpy.simulate_fmu(..., inputs = ...) .

Maybe I do not define the input array right?

input_array = np.array(np.zeros_like(time), dtype = {'names': ['time']+inputs, 'formats': ['<f8' for i in range(4)]})
for i,times in enumerate(time):
    input_array[i] = (times ,np.random.randint(1,5), np.random.randint(1,5), np.random.randint(1,5)) 

Thanks in advance!

I attached the jupyter notebook / python scripts and the fmu ( linux64) .
input_problems.zip

Hi,

Currently I export an FMU for co-simulation with the solver being Cvode, but now I would like to try a different solver. So I'm wondering what solvers are supported for co-simulation and model exchange, respectively? How can one configure the parameters for those solvers, e.g. changing tolerance?

Best regards,
Francis

For model exchange the values at time == start_time are not recorded.

from fmpy.examples.coupled_clutches import simulate_coupled_clutches

result = simulate_coupled_clutches(fmi_type='ModelExchange', show_plot=False)

print(result['time'][:5])

Output:

...
Simulating CoupledClutches.fmu (FMI 2.0, ModelExchange, CVode)...
Done.
[ 0.02  0.04  0.08  0.1   0.12]

It would be nice to be able to set the (commication) step size for CS FMUs in the GUI.
Currently the field "step size" is greyed out, and one can only enter "Max. samples" which seems to be the number of equidistant communication timesteps.

Hello,
Thanks for FMPy.
Is it possible to simulate a FMU 1.0 on Raspi. I have tried it and i get an error.

The method setInteger (FMI2) contains the implementation of setString:

fmi2.py - v0.0.8.

def setInteger(self, vr, value):
    value = map(lambda s: s.encode('utf-8'), value)
    value = (fmi2String * len(vr))(*value)
    self.fmi2SetString(self.component, vr, len(vr), value)

fmi2.py - correction

def setInteger(self, vr, value):
    vr = (fmi2ValueReference * len(vr))(*vr)
    value = (fmi2Integer * len(vr))(*value)
    self.fmi2SetInteger(self.component, vr, len(vr), value) 

I was investigating on a really strange issue today: I was using fmpy in our project together with a larger multiprocessing setup (doing parallel simulations) and I encountered a strange error (which was hard to debug due to multiprocess stuff) when running on Linux (funnily, Windows worked). The result was a BrokenProcessPool error or a queue.Full error.
After several hours of slicing away code line by code line, I ended up with the following script:

def use_fmpy(*args):
    import fmpy
    return fmpy.read_model_description('bouncingBall_me.fmu') 

if __name__ == "__main__":
    import tables  ##### TRIGGER HERE

    import concurrent.futures as cf
    with cf.ProcessPoolExecutor(max_workers=2) as pool:
        futures = []
        for i in range(10):
            future = pool.submit(use_fmpy, None)
            futures.append(future)
        cf.wait(futures)

What made it even harder was the fact that the import tables line triggered the error in the multiprocessing subprocesses. With the pytables import I got the error, without pytables, the error wasn't appearing.

So I first suspected pytables for being the source of the problem, but after a while I took a closer look at the read_model_description function. (BTW I tried out all available fmpy versions) The first lines of this function look like this:

It uses the zipfile module to read the modelDescription.xml file from the FMU file. Putting a return just before this code block made my example above work in all cases, so it must have had to do with the zipfile use.

Googling brought immediately a possible issue: It seems that the zipfile module is not threadsafe. So my idea was to replace the zipfile use by a call to unzip:

import subprocess
root = etree.fromstring(subprocess.check_output(
    ['unzip', '-p', filename, 'modelDescription.xml'])) 

And surprise: it worked! No more problems with BrokenProcessPool or queue.Full errors.

So my question here: would it be possible that you replace all the zipfile usages by something more robust? I would really love to use fmpy instead of pyfmi, but this issue cost me already 5 hours of debugging and it is currently a blocker for us.

Hi, I have a very complicated FMU model. Sometimes when I simulate the model using the template custom_input.py, it gives some errors. However, when I reopen the Spyder console, the errors disappeared. Anyway, part of the errors are as follows.

File "D:\Software\Miniconda3\envs\py36_32\lib\site-packages\spyder\utils\site\sitecustomize.py", line 705, in runfile
execfile(filename, namespace)

File "D:\Software\Miniconda3\envs\py36_32\lib\site-packages\spyder\utils\site\sitecustomize.py", line 102, in execfile
exec(compile(f.read(), filename, 'exec'), namespace)

File "D:\Software\Miniconda3\envs\py36_32\lib\site-packages\fmpy\fmi2.py", line 386, in init
super(FMU2Slave, self).init(**kwargs)

File "D:\Software\Miniconda3\envs\py36_32\lib\site-packages\fmpy\fmi2.py", line 84, in init
super(_FMU2, self).init(**kwargs)

File "D:\Software\Miniconda3\envs\py36_32\lib\site-packages\fmpy\fmi1.py", line 131, in init
self.dll = cdll.LoadLibrary(libraryPath)

File "D:\Software\Miniconda3\envs\py36_32\lib\ctypes_init_.py", line 426, in LoadLibrary
return self._dlltype(name)

File "D:\Software\Miniconda3\envs\py36_32\lib\ctypes_init_.py", line 348, in init
self._handle = _dlopen(self._name, mode)

OSError: [WinError 8] Not enough storage is available to process this command

Hi, as we are using conda for managing package dependency, fmpy gets installed as conda package with a dependency to pywin32. So far so good.

But when I run python setup.py develop in my project (which requires fmpy in its setuptools config), setuptools fails because it cannot find the dependency pypiwin32 (there is only pywin32 installed).

Any reason why the dependency is on pypiwin32 and not on pywin32 ? It seems that the original pywin32 project has been very active recently, see latest release is just 4 weeks old (https://pypi.python.org/pypi/pywin32). Would it be possible to switch back to pywin32 instead of pypiwin32 ? This would make my life much easier.

The only workaround I am currently seeing for my project is to remove fmpy from the requirements list (so that setuptools doesn't check for it) and manually make sure that fmpy and dependencies are installed correctly.

fmi2Instantiate is called with the wrong path for the fmuRessourceLocation,
since self.fmuLocation points to the unzipped FMU, but the fmuRessourceLocation should point to the resources directory of the unzipped FMU.

fmi2.py

self.component = self.fmi2Instantiate(self.instanceName.encode('utf-8'), kind, self.guid.encode('utf-8'), self.fmuLocation.encode('utf-8'), # fix e.g.: + b'/resources' byref(callbacks), visible, loggingOn)

How can I specify the values of (static) parameters in an FMU? By "parameter" I mean, causality == parameter.

Thanks!

Currently one can only specify a fixed output interval for ME simulation.
It would be beneficial to have a possibility to to also get output values at events.
So it would be best to have the following options:

• constant output interval (current implementation)
• all output values
• Constant output interval + event points

Hi,
I tried to launch FMPy on a Raspberry Pi model b. The installation was successful and I managed to load the 32bit LINUX FMU (Model Exchange). I can also display the model information. As soon as I start simulating, I get the error message that the shared object file cannot be found.

But i can see the unzipped *.so file in that folder ? Any ideas?

Mike

For the bouncingBall.fmu (64bit Windows) from the FMU-SDK 2.0.5 I get an error with FMPy when simulating (tested with the development branch for certain combinations of output_interval and stop_time.
For example for

result = simulate_fmu(fmu, output_interval=0.01, stop_time=1)

I get:
[CVODE ERROR] CVode
At t = 0.996514, mxstep steps taken before reaching tout.

....
However for

result = simulate_fmu(fmu, output_interval=0.001, stop_time=1)
There is no error.

bouncingBall.zip

• FMPy/fmpy/util.py

  Line 18 in 08a2e3c

  traj the trajectoies read from the CSV file

  trajectoies --> trajectories

• FMPy/fmpy/util.py

  Line 147 in 08a2e3c

  """ Validate a simulation result agains a reference result

  agains --> against

• FMPy/fmpy/util.py

  Line 496 in 08a2e3c

  sources_dir: directory that conatains the FMU's source code

  conatains --> contains

I have the changes in the branch https://github.com/traversaro/FMPy/tree/patch-1 , but I did not open a PR given that, citing the contributing guide:

For legal reasons we're currently not accepting pull requests. Please create an issue if you have suggestions for improvements.

When simulating a 2.0 CS FMU, the relative_tolerance argutment of simulate_fmu is not used (this parameter is only currently intended for the CVODE Solver of ME FMUs):
in

Do I understand

It would be good, if it were possible to set the tolerance for simulating a 2.0 CS FMU explicitly, perhaps be using the relative_tolerance parameter currently only used for ME FMUS with CVODE.

Is it possible to perform Global Sensitivity Analysis in Python after exporting the model from Dymola to Python using FMPy? I want to see the effect of different Parameters and the correlation between them.

Hi,

I just started learning this FMI world and FMPy seems to be the best opensource option that I could use right now. Do you have any example code of cosimulation? I plan to do it for building physics and control operations from different frameworks and would like to use FMI and FMPy to execute them.

Thanks!

What would be the recommended way to simulate interconnected FMUs using FMPy?