This example shows how a PID controller can be tuned automatically in KULI avoiding cumbersome manual tuning. For this task the cost function of the optimization problem is defined within a subsystem in KULI and allows easy adaption respectively reusage. The provided cost function penalizes control deviations and oscillations of the controlled variable. In this case the KULI model is externally controlled from Python addressing the COM interface offering a wide range of possibilities to adapt the optimization procedure to user specific needs.
To demonstrate this procedure the air temperature of a passenger compartment is controlled via the external heat supply of a point mass which represents for example a PTC heater. Several transient simulations will be conducted with different p-, i- and d-gains until the parameters for an optimal PID controller (in the sense that these minimize the cost function) are found by the python script automatically.
Keywords: step response, control theory, automatic tuningUsable from release: KULI 14