Il 02/03/2013 0.55, Lennart Ochel ha scritto:
Hej,
it is already fixed by Vitalij some seconds after my mail :-)
I love you, folks :)
Francesco
Kind Regards,
Lennart
-----Ursprüngliche Nachricht----- From: Francesco Casella
Sent: Saturday, March 2, 2013 12:27 AM
To: [email protected]
Subject: Re: Step size boundaries
Hi everyone,
As far as I know we do not support step size boundaries. The dassl
code that we are using supports that feature, so it is probably easy
to introduce that.
This would be a very useful feature for us people doing physical
modelling. In particular, being able to set a lower bound on the time
step might allow to avoid wasting CPU time to accurately compute fast
transients on a time scale which is not interesting for modeller (e.g.,
fast pressure transients in a heat exchanger network).
I've been asking this feature to Dymola developers for 10 years, but
they won't provide it, I guess because they are afraid some of their
customers might mis-use it, then complain about wrong results later.
Also being able to provide the initial step length when the integrator
is started (or re-started), overriding the default choice of the solver
can be very useful. Think of sample-data systems with a time event every
millisecond.
I think we can afford giving people the power of making their choices
regarding integration methods, if they know what they're doing :)
... obviously our helptext is invalid, because at least the following
methods are missing: radau1, radau3, radau5, lobatto2, lobatto4
May I ask that this is fixed ASAP? Only a few people can afford to spend
enough time on the source code to find out what OMC can do :)
Francesco
Kind Regards,
Lennart
-----Ursprüngliche Nachricht----- From: Christoph Höger
Sent: Friday, March 01, 2013 3:20 PM
To: [email protected]
Subject: Step size boundaries
Hi all,
I noted that the apache commons ODE solvers contain a step size boundary
(min/max stepsize) for adaptive step size integrators. Is there any such
thing for omc?
And as a related question: Are the other solvers in SimulationRuntime
supported (dopri45, radau)?
best regards,
Christoph
--
Francesco Casella - Ph.D.
Dipartimento di Elettronica, Informazione e Bioingegneria
Politecnico di Milano
Via Ponzio 34/5
I-20133 MILANO - ITALY
Tel: +39-02-2399-3465 (Leonardo)
+39-02-2399-7749 (Polo di Cremona)
Fax: +39-02-2399-3412
e-mail: [email protected]
web: http://home.dei.polimi.it/casella
Skype: callto://francesco.casella
