PID Controllers, 2nd Edition

Chapter 7.3 - Control Paradigms: Feedforward Control

Disturbances can be eliminated by feedback. With a feedback system it is, however, necessary that there be an error before the controller can take actions to eliminate disturbances. In some situations it is possible to measure disturbances before they have influenced the processes. It is then natural to try to eliminate the effects of the disturbances before they have created control errors. This control paradigm is called feedforward. The principle is simply illustrated in Figure 7.6. Feedforward can be used for both linear and nonlinear systems. It requires a mathematical model of the process.

As an illustration we will consider a linear system that has two inputs, the control variable u and the disturbance v, and one output y. The transfer function from disturbance to output is Gv, and the transfer function from the control variable to the output is Gu. The process can be described by the following equation:

07_3-1.jpg

07_3-2.jpg

Figure 7.6 Block diagram of a system with feedforward control from a measurable disturbance. where the Laplace transformed variables are denoted by capitals. The feedforward control law

07_3-3.jpg

makes the output zero for all disturbances v. The feedforward transfer function thus should be chosen as:

07_3-4.jpg

The feedforward compensator is, in general, a dynamic system. The transfer function Gff must, of course, be stable, which means that Gv must also be stable. If the processes are modeled as static systems, the feedforward compensator is also a static system. This is called static feedforward.

If the transfer functions characterizing the process are given by

07_3-5.jpg

it follows from Equation (7.3) that the feedforward transfer function is

07_3-6.jpg

In this case the feedforward compensator is a simple dynamic compensator of a lead-lag type.

Since the key idea is to cancel two signals, it is necessary that the model is reasonably accurate. A modeling error of 20% implies that only 80% of the disturbance is eliminated. Modeling errors are directly reflected in control errors. Feedforward is typically much more sensitive to modeling errors than feedback control.

Since it requires process models, feedforward is not used as much as feedback control. There are, however, many cases where a lead-lag filter, as given in Equation (7.5), or even a constant feedforward gives excellent results. The availability of adaptive techniques has drastically increased the range of applicability of feedforward. Certain standard controllers have a feedforward term. Feedforward is also easy to include in distributed control systems.

Feedback and feedforward have complementary properties. With feedback it is possible to reduce the effect of the disturbances with frequencies lower than the system bandwidth. By using feedforward we can also reduce the effects of faster disturbances. Feedback is relatively insensitive to variations in the process model while feedforward, which is used directly in a process model, is more sensitive to parameter variation. Feedback may cause instabilities while feedforward does not give rise to any stability problems. To obtain a good control system, it is desirable to combine feedback and feedforward.

UNLIMITED FREE
ACCESS
TO THE WORLD'S BEST IDEAS

SUBMIT
Already a GlobalSpec user? Log in.

This is embarrasing...

An error occurred while processing the form. Please try again in a few minutes.

Customize Your GlobalSpec Experience

Category: Motion Controllers
Finish!
Privacy Policy

This is embarrasing...

An error occurred while processing the form. Please try again in a few minutes.