Influence of interface modes on the Simulink solver

The following table indicates for each mode of the Simulink interface whether it has an influence on the Simulink solver or not.

Interface mode	Influence on the Simulink solver	Comments
SL2AMECosim	Limited	Using this mode will enforce the use of a fixed-step solver for the Simulink part (see below). Sometimes, the co-simulation can influence the method to be selected for the Simulink solver, due to the stability aspects typically encountered with sampled systems.
AME2SLCosim	Limited	Sometimes, the co-simulation can influence the method to be selected for the Simulink solver, due to the stability aspects typically encountered with sampled systems. If algebraic loops are formed by the complete coupled system, this also has an effect on the Simulink solver (see Impact of algebraic loops on the Simulink solver).
AME2SL	Yes (solver type and/or settings)	Only the Simulink solver will be used to solve the equations of both the Simcenter Amesim model and the Simulink model. So, the dynamics and/or the states introduced by the Simcenter Amesim model need to be taken into account for selecting the Simulink solver. Try using the ode15s solver in Simulink, especially for stiff systems (see Implicit or explicit Simulink solver (stiff or non-stiff systems) and Simcenter Amesim and comparable Simulink solver methods for Model Exchange modes for more details). If algebraic loops are formed by the complete coupled system, this also has an effect on the Simulink solver (see Impact of algebraic loops on the Simulink solver).

Note

The goal of co-simulation is to isolate all the solvers involved such that each of them uses the most appropriate settings for its system. This is why SL2AMECosim or AME2SLCosim usually have no influence on individual solver options. However, since the subsystems are coupled, each of them is seen as discrete by the others and the chosen communication time step can imply a modification of solver settings if instabilities arise.

• For SL2AMECosim and AME2SLCosim, if co-simulation instabilities are encountered and if adjusting the co-simulation communication interval is not sufficient, Runge-Kutta methods might be selected instead of Euler methods for the Simulink solver.

• For SL2AMECosim and AME2SLCosim, it is worth mentioning that connecting the Simcenter Amesim model to the Simulink model doesn't add any continuous state to the Simulink model. This is why these two interface modes have no consequence on the choice of a discrete or a continuous solver for Simulink and only limited influence on other solver aspects.