Correspondences between Simcenter Amesim and Simulink solver methods

This section describes correspondences between Simcenter Amesim and Simulink solver methods.

Note

Most of the principles mentioned for the Simulink solver are still valid for the Simcenter Amesim solver. However, for the Simcenter Amesim variable-step solver, most of the choices are made automatically and dynamically among 17 various integration methods. This is a major difference with Simulink since users usually do not have to worry about solver aspects in Simcenter Amesim.

• Unlike Simulink, it is possible to benefit from Simcenter Amesim's variable-step solver in a co-simulation with Simulink (both with AME2SLCosim and SL2AMECosim).

• As for Simulink, AME2SLCosim and SL2AMECosim only have a limited influence on the settings of the Simcenter Amesim solver, except when instabilities are encountered due to the coupling with the Simulink system. In cases where implicit variables (algebraic loops) are generated in the Simcenter Amesim model due to the connection with the Simulink block, adapted integration algorithms are selected accordingly (DASSL) and automatically. The following figure illustrates what we can see in the Simcenter Amesim compilation window if such an algebraic loop is created. This information can also be accessed through the State Contribution pane of the Performance Analyzer ("external constraints" mentioned in the Type column).

Figure 91: Algebraic loop on the Simcenter Amesim side

• In the Simcenter Amesim model, generated or declared implicit variables must be avoided in order to be able to select a fixed-step solver (in other words: the Simcenter Amesim model should only contain explicit state variables).

• As for Simulink, if co-simulation instabilities are encountered with Simcenter Amesim fixed-step solver, Runge-Kutta methods might be selected instead of Euler or Adams-Bashforth methods.