Dear Mr. Cameron,
>>Comenting the process 3 you say "here one offdiagonal element has a larger timedelay so it is not so clear what the best pairing is" Does it means that in this
case (real transport times) we can not use the method ?
By "the method" I assume you refer to the steps on p. 443-444.
In step 3 it says that we should prefer pairings which puts minimal
restrictions on the achievable bandwidth, that is, we should avoid
to pair on a element with a large delay. The reason is that, even though
the delay may not be a "multivariable" problem, it will make it
necassry to tune that loop slowly (when considered by iself).
So you must look at steps 1 and 3 combined, and there may be cases
where they conflict. A more careful analysis may then be required,
althogh for the case we are looking, the offdiagonal delay is 5 times
larger than the delay on the other elements, so I would probably
avoid the offdiagnal pairing.
Best regards,
Sigurd Skogestad
Received on Fri Mar 7 10:49:42 1997
This archive was generated by hypermail 2.1.8 : Tue Jan 20 2004 - 09:32:15 CET