Modern Control Engineering

(Chris Devlin) #1

Small-Gain Theorem. Consider the closed-loop system shown in Figure 10–42. In


the figure and M(s)are stable and proper transfer functions.


The small-gain theorem states that if


then this closed-loop system is stable. That is, if the norm of M(s)is smaller


than 1, this closed-loop system is stable. This theorem is an extension of the Nyquist


stability criterion.


It is important to note that the small-gain theorem gives a sufficient condition for sta-


bility. That is, a system may be stable even if it does not satisfy this theorem. However,


if a system satisfies the small-gain theorem, it is always stable.


System with Unstructured Uncertainty. In some cases an unstructured uncer-


tainty error may be considered multiplicative such that


where is the true plant dynamics and Gis the model plant dynamics. In other cases


an unstructured uncertainty error may be considered additive such that


In either case we assume that the norm of or is bounded such that


where and are positive constants.


EXAMPLE 10–14 Consider a control system with unstructured multiplicative uncertainty. We shall consider robust


stability and robust performance of the system. (A system with unstructured additive uncertain-
ty will be discussed in Problem A–10–18.)

Robust Stability. Let us define


true plant dynamics
G=model of plant dynamics
unstructured multiplicative uncertainty

We assume that is stable and its upper bound is known.We also assume that and Gare
related by

G=G(I+¢m)



G



¢m

¢m=

G



=

gm ga


7 ¢m 7 6 gm, 7 ¢a 7 6 ga


¢m ¢a


G





=G+¢a


G





G





=G( 1 +¢m)


Hq ¢(s)


7 ¢(s)M(s) (^7) q 61


¢(s)


Section 10–9 / Robust Control Systems 809

(s)

M(s)

Figure 10–42
Closed-loop system.

Free download pdf