Tuning of PID Controllers for Linear Dynamic Systems Using “Brute Force and Visual” Method

Abstract

For PID controllers in control loops, which in practice still prevails, it is necessary to determine controller parameter values which assure sufficient quality and robustness of control. A lot of PID tuning methods is described and well established in daily use. They are often based either on behavior of closed control loop or on a mathematic model of the controlled plant (and its L-transfer function), which allow to determine controller parameters by direct calculation. With increasing computing power of HW and SW aids it could be beneficial to use a method for determining the optimal setting of PID controller by simulation of control loop behavior for various meaningful combinations of controller parameter values (state space search). Proposed method “brute force and visual” of PID controller tuning is based on repeated simulation of control process as a reaction to Heaviside step of desired value and disturbing value for varying parameters of controller’s P, I, D components. For discrete controller also impact of sampling period T was considered. For each of the individual simulation runs there is kept a diagram as an image file as well as the control optimality criterion value. After execution of all the simulation runs for parameter values from the considered searched state space the simulation outputs are sorted by value of the selected control optimality criterion. The parameters of the individual simulation which has scored best are used for setting of the real controller and evaluated in the real control loop. From the simulations available for wide combination of parameter values it is possible to estimate position of isles of stability and choose the PID controller values by an expert choice.