10.1 Background

Thus far in this textbook we have emphasized control-system design for single input–single output (SISO) processes; that is, processes with one output (measured) variable and one manipulated variable. We have also focused on setpoint responses, primarily because it is then easier to tune a feedback controller. In particular, it is easier to tune on-line for setpoint changes because we do not know when a disturbance is going to enter a system. We have developed a number of control-system design procedures that are based on a desired response to a setpoint change.

In practice, disturbance rejection is very important. The primary disadvantage to feedback-only control is that a disturbance must be "felt" by the output variable before there is a control-system response. The purpose of this chapter is to show how to use multiple measurements to improve the response to a disturbance. In Sections 10.2 through 10.5, we study cascade control. In cascade control, multiple output measurements are used to improve the response of the most important (primary) output to a disturbance. In the last part of the chapter (Sections 10.6 to 10.8) we cover feed-forward control, where the measurement of a disturbance is directly used to improve the system response to the disturbance. In Section 10.10 we combine feed-forward and cascade control.