A unified anti-windup strategy for SISO discrete dead-time compensators

This paper presents the development of a unified anti-windup strategy for dead-time compensators that copes with stable, integrating, and unstable single-input and single-output linear plants subject to sector-bounded and memoryless input constraints (such as magnitude or rate saturation, the most common cases in practical applications). A discrete-time Filtered Smith Predictor is used as the base control structure and in this specific case it is shown that stability conditions are related to the primary controller and fast model of the plant instead of the whole control structure. As the design is driven directly in the z-domain, implementation of the strategy using a digital controller is immediate. In addition, simulations and a practical experiment using a vapor-compression refrigeration system with time delay are provided and the results show that the strategy has good performance, close to the one presented by optimal model predictive control structures, with the great advantage of requiring no optimization problem to be numerically solved.