Modeling

You can rebuild the behaviour of a real system from function blocks in Simulink. For this you use blocks with standardised behaviour. In practice, modelling real systems in Simulink is often lengthy, error-prone and difficult. The result often only represents the basic behaviour, which deviates from real behaviour at some operating points. For reasons of time, we do not go further into modelling real systems.

Once a system has been modelled in Simulink, you can vary the controller parameters in the simulation until you are satisfied with the behaviour of the controlled system. Then you apply the controller parameters to the controller of the real system and verify whether the behaviour also fits the real system. You keep readjusting the behaviour until you have achieved all goals on the real system. The controller parameters from the simulation are often a good starting point for this.

This method is very laborious and time-consuming. You constantly have to carry out measurements on the real system and compare them with the results of the simulation. A model quickly becomes very complex.

But once you have a good model, you have many advantages: you can try out different controllers very flexibly and have the computer calculate the result directly. You do not have to measure on the real system to get results; that is easier and faster in the simulation. Nothing can break in a simulation if you set the wrong controller parameters. You can easily test planned changes to the system in advance by implementing them in the model first. Simulation is very helpful, but the path to a good model is hard. The way the system works must be known very precisely.