Repository navigation
model reduction by residualization not implemented or not documented #1168
Description
Activity
From the
model_reductiondocstring:Two methods of state reduction are possible: 'truncate' removes the states marked for elimination, while 'matchdc' replaces the eliminated states with their equilibrium values (thereby keeping the input/output gain unchanged at zero frequency ["DC"]).From the model_reduction docstring: ...
ok that smells like the residualization described by skogestad, but i can't tell for sure (would be useful to mention explicitly in docs :D for those of us not as fluent in the possibilities and standard approaches)
My bad again. I was reading docs here, which do not contain as descriptive documentation. It must have improved since then. Searching for the docstring you quoted i found that there is a "latest" url for readthedocs.
You should get a warning if you reference an older version of the documentation (see here).
An open issue for anyone reading this and having some time to contribute: while 'matchdc' is described in
model_reduction, it is not described inbalanced reduction, where it is also used.You should get a warning if you reference an older version of the documentation (see #1140 (comment)).
sure enough - i refreshed the page and there is a tooltip that shows up for 4 seconds when the page is refreshed. sounds like a readthedocs problem, but the warning should probably not go away.
FYI, I've got a branch of small documentation changes and I have updated the documentation for
balanced_reductionthere, so this will get addressed for 0.10.3.
the documentation does not explain what exactly "matchdc" does. is it residualization or something else? It seems to not just be a gain adjustment.