n-1 security

[birgits]

We currently consider n-1 security by using allowed line loading of 50% in the load case. For grid dimensioning using worst case this should be okay, as there is no feed-in in the load case. However, when using time series the feed-in and load situations in the different feeders are quite diverse and assuming a load case when the residual load in the whole grid is positive can lead to dimensioning of a feeder for n-1 security if though feed-in is always higher than load. We should either determine whether it is a load case individually for each feeder/line segment or use faulty condition calculation (see also discussion here).

• [ ] How is it done by others using time series for grid dimensioning
• [ ] Do we only need n-1 security in rings or also for stubs?
• [ ] ..

[birgits]

In #423 I started an implementation to fix this issue.