SmartDS Scenarios Issues

[adam-morse]

Hi Paulo,

First thank you for everything you are doing with DSS-Extensions. It's been an amazing experience using the tool.

I'm having an issue running high and extreme PV SmartDS, time series scenarios using DSSExtensions that I can't seem to troubleshoot. I didn't get resolution through a github issue at SmartDS:
openEDI/documentation#38

I can't expect you to download the SmartDS dataset to reproduce the issue but you can look at the files I'm working with:
https://data.openei.org/s3_viewer?bucket=oedi-data-lake&prefix=SMART-DS%2Fv1.0%2F2018%2FGSO%2Furban-suburban%2Fscenarios%2Fsolar_high_batteries_none_timeseries%2Fopendss%2Fuhs0_1247%2Fuhs0_1247--udt12274%2F

The main reason I thought it might related to InvControl elements is that the none, low and medium PV scenarios do not have any InvControl, but I really don't have a clue not being able to decipher the error code.

Any help is much appreciated,
Adam

DSSException Traceback (most recent call last)
Cell In[229], line 1
----> 1 ckt(f'Redirect {master_file}')

File ~\anaconda3\envs\holoviz_env\Lib\site-packages\altdss\AltDSS.py:450, in AltDSS.call(self, cmds)
426 def call(self, cmds: Union[AnyStr, List[AnyStr]]):
427 '''
428 Pass either a single string (with either one or multiples commands, separated by new lines),
429 or a list of strings.
(...)
448 ```
449 '''
--> 450 self.Commands(cmds)

File ~\anaconda3\envs\holoviz_env\Lib\site-packages\altdss\AltDSS.py:411, in AltDSS.Commands(self, Value)
408 if not isinstance(Value, bytes):
409 Value = Value.encode(self._api_util.codec)
--> 411 self._check_for_error(self._lib.Text_CommandBlock(Value))
412 else:
413 self._check_for_error(self._set_string_array(self._lib.Text_CommandArray, Value))

File ~\anaconda3\envs\holoviz_env\Lib\site-packages\dss_cffi_api_util.py:267, in Base._check_for_error(self, result)
265 error_num = self._errorPtr[0]
266 self._errorPtr[0] = 0
--> 267 raise DSSException(error_num, self._get_string(self._lib.Error_Get_Description()))
269 return result

DSSException: (#482) Error Encountered in Solve: Access violation
(..\SmartDS_GSO\urban-suburban\scenarios\solar_extreme_batteries_none_timeseries\opendss\uhs0_1247\uhs0_1247--udt12274\Master.dss", line: 23]

[PMeira]

Hi, Adam,

I'm having an issue running high and extreme PV SmartDS, time series scenarios using DSSExtensions that I can't seem to troubleshoot. I didn't get resolution through a github issue at SmartDS:
openEDI/documentation#38

Which engine version are you using?

I can't expect you to download the SmartDS dataset to reproduce the issue but you can look at the files I'm working with (...)

Thanks for the link, I downloaded that folder (and the solar_extreme variation), and the profiles.

I got the following error message on Linux with the latest stable release, so invalid properties as expected:

(#110) Unknown parameter "kvarlimit" (value "1.8333333333333335") for object "PVSystem.pv_load_p1ulv44811"
[file: "/tmp/smartds/scenarios/solar_high_batteries_none_timeseries/opendss/uhs0_1247/uhs0_1247--udt12274/PVSystems.dss", line: 568]
[file: "/tmp/smartds/scenarios/solar_high_batteries_none_timeseries/opendss/uhs0_1247/uhs0_1247--udt12274/Master.dss", line: 11]

From your comment in openEDI/documentation#38:

There is a note in what you cited about LegacyModels no longer being supported beyond DSS C-API v0.13.0. The earliest version of opendssdirect I was able to install this morning was 0.7.0. When I try installing earlier versions I get installation errors that are related to dss_python distribution version 0.10.7. I did get the same LegacyModels error for version 0.7.0...

LegacyModels was removed in March 2023, opendssdirect<=0.7.0 (DSS C-API version 0.12.1) should work. Not that I recommend using old versions, and I didn't wait for it to finish, but with ODD.py 0.7.0, on Linux, it seems to be running fine after enabling LegacyModels. Note that you should be able to use recent Python releases, e.g. Python 3.13 still runs fine with old releases from DSS-Extensions. LegacyModels was never available in the official OpenDSS, only in DSS C-API/AltDSS.

Updating the .DSS files as discussed in https://sourceforge.net/p/electricdss/discussion/861976/thread/4a6aa2199f/ might work. The PDF I mention was removed (merged to the online docs), but the last version is on https://sourceforge.net/p/electricdss/code/3706/tree/trunk/Version8/Distrib/Doc/PVSystem_InvControl.pdf

I did try running replacing, in PVSystems.dss:

• kvarlimit with kvarmax
• VV_RefReactivePower=VARMAX_VARS with Mode=VOLTVAR RefReactivePower=VARMAX (you might need to adjust other properties). I noticed you mentioned part of this in the openEDI post, but probably need to set Mode to VoltVar too.
• VV_RefReactivePower=VARMAX_VARS with RefReactivePower=VARMAX (you might need to adjust other properties).

With that, I was able to run without LegacyModels in the old versions, and the latest release. I waited for the command to finish on the latest release with the solar_high_batteries_none... scenario, it didn't crash.

I didn't replace PVSystemlist=[ yet, but that's deprecated and used to crash some versions of the official OpenDSS. The new property is DERList, so if all PVSystemlist contains only one element, replacing PVSystemlist= with DERList=[PVSystem. should work. See also https://dss-extensions.org/dss-format/InvControl.html

It would be a good idea to compare the results with the old LegacyModels flag (or an even older engine version, before 2020), and an updated dataset with the replacements above. In case they're enough, it might be worth to document it somewhere, maybe even add a preprocessing script to adjust the files.

Note that I don't expect the results to match exactly, there are lots of changes since the old releases besides the InvControl/Storage/PVSystem models.

[PMeira]

@adam-morse, I can reproduce your issue on Windows! Good news is that the latest dev branch (not public yet) already fixes it, and I believe the next release of the official OpenDSS does too.

[adam-morse]

@PMeira You made my day! Interesting that it works on Linux but not Windows. NREL saved their simulation results in the dataset to 'analysis' folders so I should be able to cross check against what they got in ~2020. I'll look forward to the the dev branch being released.