Teacher: José Luis Guzman

Program: Feedforward control to deal with load disturbances is a very old topic with more than 100 years of history. Most industrial processes are affected by disturbances and feedback can be inefficient to reject them because of the reactive behavior of the control loop. Feedforward control approaches have become like a standard and simple solution to improve the disturbance rejection effect when disturbances are measurable, since the feedforward control contribution is injected before affecting the process output. In fact, under perfect situations, the disturbance effect can be completely removed.

Nevertheless, those perfect situations are seldom possible and the feedforward is not realizable in most occasions. In those cases, static feedforward compensators are typically used in industry, but with this solution non-desirable overshoots are obtained and the performance is not the expected one. So, could the feedforward compensator be tuned in those cases in order to improve the disturbance rejection response? The answer is, yes.

This course motivates why feedforward compensators should be tuned and how different tuning rules can be proposed. First, it is shown that when inversion problems arise, the feedforward compensator should not be designed in open loop anymore. Now, a connection between the feedback controller and the feedforward compensator is required. This analysis will lead to two different groups of tuning rules based on closed-loop or open-loop designs. So, this course summarizes these two groups of tuning rules as a result of the research on this topic during the last 10 years. Numerical simulation results and experimental tests on the greenhouse temperature control problem will be shown to illustrate the main contributions.

Duration (hr):  2

Language:  English