different output dimension of different scenarios with the same ss.TDS.config.tf=10

[xiaojiangsi]

I used andes loop to generate data for time domain simulation under different faults and got different dimensional results, if I follow 0.033 steps it should be 30827, but the final data appears in multiple cases. Why did this happen?
for example :

it cleared the fault at time 1.2, and the picture shows the part of the time series.
but when load level changed, the time series is as follow:

and somehow the dimension of the output is bigger than 30827.

[cuihantao]

The trapezoidal method in ANDES uses a large step size by default. If nonconvergence for a step happpens, it will heuristically reduce the step size. When convergence improves, it will adjust the step size back to TDS.config.tstep. You can turn off the automatic shrinking of step size (see docs for TDS.config.shrinkt), but your first case will likely fail after t=1.0001.

If you need uniform step size, one way is to post-process the data and interpolate at the desired time.

The second solution is to use a small step size (1/120, e.g.) and turn off shrinkt. This will slow down simulations across the board. Most cases should be able to complete, but there is no guarantee of convergence for all cases.

[jinningwang]

Hello @xiaojiangsi ,

Hope you find a helpful answer here. Please MARK OUT an answer to your question if you find any reply appropriate. In this way, the Q&A will be a valuable input to this community that can be helpful for future users.
We commit efforts to maintain this open-source software. We will be grateful if you can contribute to it if our work really helps you somewhere.

Regards,
Jinning