Adding an "FLoad" dynamic load causing error

[MaKla89]

For a very small test-grid (4 bus, 1 slack, 1 synGen, 2 lines, 2 transformers, ONE load) I tried to go from static load (PQ) to dynamic load (FLoad). Yet, even if I keep the FLoad inactive (u = 0), running TDS raises this error:

ValueError: All arrays must be of the same length

The following screenshot contains the FLoad-settings and the Traceback:

As there is very little to read in the FLoad-documentation, I am not sure whether I'm 100% blind or whether something is missing :/

When I delete Row 1 of the xlsx file, then the model is running fine again (obviously without the dynamic load, but it works).

[jinningwang]

Hi @MaKla89, it's great to see you here again! It seems like there might be a bug. Would you mind sharing the relevant case file and minimal code with me? I'd be happy to take a look when I have some time.

[cuihantao]

Looks like cases/ieee14/ieee14_fload.json can replicate it.

It was broken somewhere between version 1.5.0 and 1.6.0. Doing a git bisect can identify the commit with issue.

[MaKla89]

Hi @MaKla89, it's great to see you here again! It seems like there might be a bug. Would you mind sharing the relevant case file and minimal code with me? I'd be happy to take a look when I have some time.

Sure thing, thanks! :) Here you go, that system is so small, that it can directly be used as a minimal case:
v_coll_sys.xlsx

Here the minimal-code:
FLoad problem.txt

[jinningwang]

The version test for 1.5.2 failed, but the test was successful for version 1.4.4.

Minimal code:

ss = andes.load(andes.get_case('ieee14/ieee14_fload.json'))
ss.PFlow.run()
ss.TDS.init()

[jinningwang]

It seems to be a quick fix and I'm working on this. Expect to solve this in 1-2 days.

[MaKla89]

Excellent news, thanks for the great support and communication :)

[jinningwang]

Using 1.4.4 can be a possible option before this was fixed, I don't have a clear timeline to fix this.

[jinningwang]

Update: Dr. Cui has fixed this, see #444

[MaKla89]

Perfect, I'll try it right away and mark this as the answer to my question :)

[jinningwang]

@cuihantao Hantao, I'm having trouble figuring out how to fix this error. Would you be able to take a look at it when you have some free time?

Error

ANDES version: 1.5.1.post2+ge8465ce0, case: ieee14_fload.json

The error occurred in method ModelCache.refresh in model.py. The error was caused by the bus value being none by ss.FLoad.as_dict(vin=True). Output results are shown below:

ss.FLoad.as_dict(vin=False)

{'uid': array([0]),
...
 'bus': [2]}

ss.FLoad.as_dict(vin=True)

{'uid': array([0]),
...
 'bus': array([], dtype=float64)}

First bad commit

After inspection, the issue first occurred in commit e8465ce. The procedures to identify it are listed below:

1. Use git bisect to locate the first known issue commit.
2. Start the git bisect by running git bisect.
3. Mark the first known good commit 9c2a5c8 using git bisect good 9c2a5c84.
4. Mark the first known bad commit f417cf2 using git bisect bad f417cf2e.
5. Perform the bisection method and test run, and mark each as good or bad, by running the following commands:

git checkout <commit>
pip install -e .
andes prep
andes st
andes run <andes_path>/andes/cases/ieee14/ieee14_fload.json -r tds

Finally, the first bad commit can be identified.