Page 490 - Mechanical Engineers' Handbook (Volume 2)
P. 490

6 Stability  481

            6.2 Polar Plots
                           The frequency response of a system described by a transfer function G(s) is given by
                                                 y (t)   X G(j )  sin[ t   G(j  )]
                                                                      /
                                                                           1
                                                  ss
                           where X sin  t is the forcing input and y (t) is the steady-state output. A great deal of
                                                             ss
                           information about the dynamic response of the system can be obtained from a knowledge
                           of G( j ). The complex quantity G( j ) is typically characterized by its frequency-dependent
                           magnitude  G( j )  and the phase G( j ).  There are many different ways in which this
                                                       /
                           information is handled. The manner in which the phasor G( j ) traverses the G( j )-plane as
                           a function of   is represented by a polar plot. This is represented in Fig. 28. The phasor OA
                                                                     /
                           corresponds to a magnitude  G( j  )  and the phase G( j  ).  The locus of point A as a
                                                       1
                                                                          1
                           function of the frequency is the polar plot. This representation is useful in stating the Nyquist
                           stability criterion.
            6.3  Nyquist Stability Criterion
                           This is a method by which the closed-loop stability of a LTI system can be predicted by
                           knowing the frequency response of the 1oop transfer function. A typical CLCE may be
                           written as

                                                          1   GH   0
                           where GH is referred to as the 1oop transfer function.
                                                      2
                              The Nyquist stability criterion is based on a mapping of the so-called Nyquist contour
                           by the loop transfer function. Figure 29a shows the Nyquist contour, and Fig. 29b shows its
                           typical mapped form in the GH plane. Let C be the contour in the s-plane to be mapped.
                                                               1
                           Utilizing this information and the poles of GH, the Nyquist stability criterion can be stated
                           as follows: A closed-loop system is stable if and only if the mapping of C in the GH plane
                                                                                     1
                           encircles the ( 1,0) point N number of times in the counterclockwise direction, where N is
                           the number of poles of GH with positive real parts.
                              Clearly, if GH does not have any poles in the right-half plane, then C should not encircle
                                                                                   1
                           the ( 1,0) point if the closed-loop system is to be stable.
                              If there are poles of GH on the imaginary axis, then the Nyquist contour should be
                           modified with appropriate indentations, as shown in Fig. 30. An example is given to map
                           such an indented contour with the corresponding loop transfer function. Consider



















                                                                Figure 28 Polar plot.
   485   486   487   488   489   490   491   492   493   494   495