Dead Time Compensators

Dead time compensators were among the first predictive controllers formulated. This is not surprising since for time-delay systems, prediction of the behaviour of the system in the future is necessary to formulate the appropriate control action. For a non-predictive controller, the control action or the effect of control action cannot be realised until some time in the future when the system variables may have changed considerably such that the control action is no longer appropriate. While the proportional-integralderivative (PID) controller is widely used in industry, its performance when used in time-delay systems is known to be limited. If a time-delay system is controlled by a PID controller, the derivative action is often disabled to reduce excessive overshoot and oscillations, due largely to misinterpretation of the non-responsiveness of the system (Hagglund and Astrom, 1991). On the other hand, with no derivative action, there is no predictive capability in the remaining PI control action, which is precisely what is needed for these systems. Dead time compensators are employed when better performance is demanded.