Anti-windup schemes comparison for digital repetitive control

Abstract

As other Internal Model Principle based strategies, digital repetitive control uses an internal model which provides infinite gain at specific frequencies. In systems subject to actuator saturation, a controller with this characteristics is highly prone to windup effect. Since linear design does not consider the actuator saturation this non linear behaviour may compromise the performance and even the stability of the system. One way to deal with this problem is to avoid saturation by selecting an actuator with larger capacity but it increases the implementation cost. In other cases, an anti-windup scheme is necessary. The main goal of the anti-windup strategy is threefold: to obtain a faster recovery of the system after saturation, to achieve less performance degradation and to preserve stability. In this paper three different anti-windup schemes have been selected from the available literature to address the windup problem in digital repetitive control. The design and implementation issues are discussed. A simulation example compares the results when saturation is reached either during the transient response or in steady state.