The effect of derivative action

The derivative action in a PID controller changes the control action in proportion to the size of the derivative of the error. Since it works on the derivative of the error, derivative action is completely unable to control a process on its own (if the error doesn't change derivative won't do anything). As a result derivative action is always used in conjunction with one or more of the other control modes. The most common combinations are PID and PD (this is more common in guidance systems, where the process itself contains an integrator).

Derivative action is impossible to examine analytically with the tools we have available in this course. We can however look at the effects derivative action can have on a controlled response through simulation.

A movie presentation of what derivative action does is here. The model used in the demonstration is here.

Derivative action acts to stabilise the controlled response, allowing higher gains and shorter integral time constants to be used, and this results in much tighter control. Another effect of adding derivative action is that overshoots on controlled responses can be restricted.

Although, in theory, derivative action can always be used to provide improvement in controlled responses it is in fact not commonly used in the process industries. The reason for this is the problems that can result when using derivative action with noisy measurements. A movie demonstrating some of these effects can be found here.