Use metric instead of growthrate for vertical stability

[bclyons12]

No description provided.

[TimSlendebroek]

@bclyons12 can you push changes in the data_struct_extra.json as well?

[bclyons12]

@TimSlendebroek fixed with 702dfd4

@orso82 ready to merge I think

[orso82]

I am ok with having a generalized stability metric, but should we at least have a convention that >1 is unstable or something like that?

[bclyons12]

We can call it something other than metric, if you prefer, but I don't think we can have a standard convention like that. For vertical stability, one of the metrics is supposed to be greater than 0.15 (probably strictly greater than 0, but they add a margin). The other is supposed less than 10. In ActorVerticalStability, I just put the definition and desired limits in the description.

[orso82]

I feel like we should add such operational limits to the stability IDS, which is designed to contain various operational limits and express them as a fraction of threshold. Please take a look at the ActrorPlasmaLimits and see if this could work.

IMASDD.stability  Stability of the current scenario
├─ all_cleared  True (1) if all stability limits are cleared, otherwise False (0)
├─ collection[:]  Limits data structure
│  └─ identifier  Limit collection identifier
│     ├─ description  Verbose description
│     ├─ index  Integer identifier (enumeration index within a list). Private identifier values must be indicated by a negative
│     │  index.
│     └─ name  Short string identifier
├─ model[:]  Limits data structure
│  ├─ cleared  True (1) if a given stability limit is cleared, otherwise False (0)
│  ├─ fraction  Fraction of limit 
│  └─ identifier  Limit model identifier
│     ├─ description  Verbose description
│     ├─ index  Integer identifier (enumeration index within a list). Private identifier values must be indicated by a negative
│     │  index.
│     └─ name  Short string identifier
└─ time [s]  Generic time

[bclyons12]

@orso82 I'm just not sure a fraction of a threshold is all that useful. What does that mean for growth rate, for example, which can be negative? And then a fraction doesn't necessarily capture the importance of the metric. Something might be fine at 90% of a threshold. Others you might want orders of magnitude below. To me it seems better to just report raw numbers in IMAS and let the user decide in the application how to evaluate those.

[orso82]

Point taken @bclyons12

I agree with you that with the current approach information is lost.

It would be best if we stored:

1. the value (time dependent)
2. the threshold (time dependent)
3. the direction of the inequality operation > or < (this can be done with an integer +1 for > and -1 for <)

This will allow to reconstruct fraction if/when needed