[bug] PowerFactory export validation fails when number of switches is different between PF and PP

[Heiner92]

Bug report checklis

• Searched the issues page for similar reports

• Read the relevant sections of the documentation

• Browse the tutorials and tests for usefull code snippets and examples of use

• Reproduced the issue after updating with pip install --upgrade pandapower (or git pull)

• Tried basic troubleshooting (if a bug/error) like restarting the interpreter and checking the pythonpath

Reproducible Example

I used the PF exporter on a meshed low voltage grid.

Issue Description and Traceback

The PowerFactory export GUI is a handy tool. Unfortunately there is a bug in the validation part of the exporter.
I used the following settings:

The following output get displayed in the PF console:

set 5 lines out of service
set 0 impedances out of service
running pandapower loadflow
pandapower net converged: True
Error while verifying net: Unalignable boolean Series provided as indexer (index of the boolean Series and of the indexed object do not match).
Traceback (most recent call last):
File "C:\Users\Heiner.Frueh\Documents\Test\Lib\site-packages\pandapower\converter\powerfactory\main_pf.py", line 195, in calc
run_verify(net)
File "C:\Users\Heiner.Frueh\Documents\Test\Lib\site-packages\pandapower\converter\powerfactory\main_pf.py", line 134, in run_verify
all_diffs = validate_pf_conversion(net, ** load_flow_params)

File "C:\Users\Heiner.Frueh\Documents\Test\Lib\site-packages\pandapower\converter\powerfactory\validate.py", line 269, in validate_pf_conversion
wrong_switches = net.res_switch. loc[

File "C:\Python311\Lib\site-packages\pandas\core\indexing.py", line 1153, in _getitem_
return self ._ getitem_axis(maybe_callable, axis=axis)

File "C:\Python311\Lib\site-packages\pandas\core\indexing.py", line 1375, in _getitem_axis
return self ._ getbool_axis(key, axis=axis)

File "C:\Python311\Lib\site-packages\pandas\core\indexing.py", line 1171, in _getbool_axis
key = check_bool_indexer(labels, key)

File "C:\Python311\Lib\site-packages\pandas\core\indexing.py", line 2575, in check_bool_indexer
raise IndexingError(
pandas.errors. IndexingError: Unalignable boolean Series provided as indexer (index of the boolean Series and of the indexed object do not match).
Python-Skript 'convert2pandapower' erfolgreich ausgeführt

For some unknown reason the number of switches is different between PF and PP. The bug is caused by lines 268-271 in converter/powerfactory/validate.py:

pf_closed = pf_results['pf_switch_status']
wrong_switches = net.res_switch.loc[
    pf_closed != net.switch.loc[pf_closed.index, 'closed']
].index.values if 'res switch' in net.keys() else []

I fixed it by changing it into:

pf_closed = pf_results['pf_switch_status']
wrong_switches = pf_closed.loc[
    pf_closed != net.switch.loc[pf_closed.index, 'closed']
].index.values if 'res switch' in net.keys() else []

Expected Behavior

I can see the rest of the validation output.

Installed Versions

DlgSILENT PowerFactory 2023
python: 3.11
pandapower: 2.14.6

<style> </style>

os	Windows
OS-release	10
Version	10.0.19045
machine	AMD64
processor	Intel64 Family 6 Model 85 Stepping 7, GenuineIntel
byteorder	little
LC_ALL	None
LANG	None
LOCALE	de_DE. cp1252
pandas	02.01.2002
numpy	1.25.2
pytz	2023.3. post1
dateutil	02.08.2002
setuptools	65.5.0
pip	23.02.2001
Cython	3.0.5
'pytest	07.04.2003
hypothesis	6.88.3
sphinx	07.02.2006
blosc	01.11.2001
feather	0.4.1
xlsxwriter	03.01.2009
1xml. etree	04.09.2003
htm151ib	01. Jan
pymysq1	None
psycopg2	None
jinja2	03.01.2002
IPython	8.17.2
pandas_datareader	None
bs4	04.12.2002
bottleneck	01.03.2007
dataframe-api-compat	None
fastparquet	01.10.2023
fsspec	2023.10.0
gcsfs	None
'matplotlib	03.08.2001
numba	0.58.1
numexpr	02.08.2007
odfpy	None
openpyx1	03.01.2002
pandas_gbq	None
pyarrow	14.0.0
pyreadstat	None
pyx1sb	None
s3fs	None
scipy	01.11.2003
sqlalchemy	2.0.23
'tables	03.09.2001
tabulate	0.9.0
xarray	01.10.2023
x1rd	2.0.1
zstandard	0.22.0
tzdata	2023.3
qtpy	02.04.2001
pyqt5	None

Label

• Relevant labels are selected

[rbolgaryn]

Thank you! I think it could be due to the exporter replacing low-impedance lines by switches, creating new switches. I will fix it.

[Heiner92]

Thank you very much! As far as I know this bug was also present in the previous version, where the option to replace low-impedance lines with switches was not yet present (at least not in the GUI).

Could you please also add an option to specify the accuracy of the validation output or the units (e.g. kVA instead of MVA), or make it dependet on the lowest nominal voltage? Our use case includes LV-grids, which are tricky to model in PF because of the many scaling factors that may or may not be present. Hence we would like to be able to see any differences in loadflow results after the export. Currently the validation is displayed with 0.1 MVA accuracy. I could also make a PR sometime next week for you to decide.

In case you need a pfd grid for testing, please get in contact with me :)