Investigation of P and PD Controllers' Performance in Control Systems with Steady-State Error Compensation

摘要

A proportional-integral-derivative (PID) controller is the most common controller in industrial process control systems [1]. The three-term control action takes into account the present error (P), accumulation of past errors (I) and prediction of future errors (D), and the weighted sum of these actions depends on the weighting coefficients that are adjusted to the optimum values for the desired response of controlled process. Particular applications require only one (P, I) or two (PI, PD) control actions to provide the required accuracy of control. Choosing of control law depends on requirements to a control system. One of frequent requirements is absence of steady-state error between the set-point and the controlled output. The steady-state error is an inbuilt characteristic of feed-back control systems with P and PD controllers and the most common way of the error elimination is to add an integral term (I) to the control law. Although the integral term of PI and PID control laws allows automatic elimination of the error, it tends to increase the oscillatory or rolling behavior of the process response. The other disadvantage related to the integral term is an integral windup problem that occurs at large set-point changes and the process saturation [1-3].