Introduction

The figure below shows in more detail how the interfaces work. In the standard interface the Simcenter Amesim part of the system gets state variables and input variables from Simulink and calculates state derivatives and output variables. The process of exchanging this information is controlled entirely by the Simulink solver. In this case one could say that we import the equations into Simulink.

In the co-simulation case, the only variables exchanged are the input and output variables. In this case, the rate of exchange is set according to a user-defined parameter. As the name indicates, the model is not entirely in the hands of a single piece of software (Simulink) but it is a co-operation between two (or more) software packages. It is important to realize that by exchanging only input and output variables at a certain sample rate there is a loss of information.

Figure 28: The two Simcenter Amesim-Simulink interfaces, exchange of information

This can be compared with the difference between a continuous and a sampled controller. Normally, the smaller the sample rate used, the closer to the continuous result we get. Another problem is the loss of information about possible cross couplings between the systems since we do not communicate information about states and state derivatives.

Warning

All these factors mean that co-simulation can be difficult to set up if the two separate systems are continuous. In this case you should try to find an interface between the systems where the coupling is as weak as possible.

The obvious situation where co-simulation may be used is of course when the interface between the systems is sampled, for example when using a sampled controller.