You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Reported by otter on 5 Jan 2015 20:59 UTC
Simulating Blocks.Nonlinear.PadeDelay with delayTime=0.001 and n=4 is practically halting the simulation (no progress in simulation time). Linearizing this model give a completely wrong linear model (number of zeros/poles are wrong).
Comment by otter on 5 Jan 2015 21:09 UTC
Fixed in 187f8b3 (in trunk) by a more reliable numerical implementation and adding 2 test models in ModelicaTest:
A new parameter balance=false/true introduced. The default balance=false results in the previous implementation of MSL 3.2.1 (for backwards compatibility). For balance=true, the state space representation is balanced, resulting in a much smaller norm of the system matrix. Simulation of the critical system does now pose not any problems. Also higher order, say n=10, can now be simulated without problems. Furthermore, linearization results in the correct linear system.
In r7840 the Bode plot computation in Modelica_LinearSystems2 was improved to handle systems with large positive zeros in a numerically reliable way. With this change, the Bode plot of the critical system, also with higher n, is now also correct.
Reported by otter on 5 Jan 2015 20:59 UTC
Simulating Blocks.Nonlinear.PadeDelay with delayTime=0.001 and n=4 is practically halting the simulation (no progress in simulation time). Linearizing this model give a completely wrong linear model (number of zeros/poles are wrong).
Migrated-From: https://trac.modelica.org/Modelica/ticket/1637
The text was updated successfully, but these errors were encountered: