Page 834 - The Mechatronics Handbook
P. 834

066_Frame_C26  Page 19  Wednesday, January 9, 2002  1:59 PM









                       By combining (26.24) and (26.25), it can be proven that the feedback system has no roots in the closed
                       right half plane when h < π/2. Furthermore, the system is unstable if h ≥ π/2. In particular, for h = p/2
                       there are two roots on the imaginary axis, at ± j1. It is also easy to show that, for any h > 0 as k → ∞,
                       the roots converge to


                                                   r k →  1   2kp 
                                                              ---------   j2kp±
                                                        --- −ln
                                                        h      h 
                       As h → 0, the magnitude of the roots converge to ∞.
                         As illustrated by the above example, property (iii) implies that for any given real number s there are
                       only finitely many r k ’s in the region of the complex plane

                                                       =
                                                      s : { s ∈  : Re s() ≥ σ}

                       In particular, with σ = 0, this means that the quasi-polynomial χ(s) can have only finitely many roots in
                       the right half plane. Since the effect of the closed-loop system poles that have very large negative real
                       parts is negligible (as far as closed-loop systems’ input–output behavior is concerned), only finitely many
                       “dominant” roots r k , for k = 1,…,m, should be computed for all practical purposes.

                       Dominant Roots of a Quasi-Polynomial

                       Now we discuss the following problem: given N(s), D(s), and h ≥ 0 , find the dominant roots of the quasi-
                       polynomial

                                                     χ s() =  Ds() + e Ns()
                                                                  −hs

                                                                                                     ,
                       For each fixed h > 0, it can be shown that there exists σ max  such that χ(s) has no roots in the region   s
                                                                                                    max
                       see [11] for a simple algorithm to estimate σ max , based on Nyquist criterion. Given h > 0 and a region
                       of the complex plane defined by σ min  ≤ Re(s) ≤ s max , the problem is to find the roots of χ(s) in this region.
                         Clearly, a point r = σ + jω in   is a root of χ(s) if and only if

                                                D s +  jw) =  e –  −hs −jhω N s +  jw)
                                                                     (
                                                  (
                                                                e
                       Taking the magnitude square of both sides of the above equation, χ(r) = 0 implies

                                      A s x() : D s +(=  x)D s –  x) e −2hs N s +  x)N s –  x) =  0
                                                       (
                                                                             (
                                                                     (
                                                              –
                       where x = jω. The term D(s + x) stands for the function D(s) evaluated at σ + x. The other terms of
                       A σ (x) are calculated similarly. For each fixed σ, the function A σ (x) is a polynomial in the variable x. By
                       symmetry, if x is a zero of A σ  , then (−x) is also a zero.
                         If A σ (x) has a root x    whose real part is zero, set r    = s + x   . Next, evaluate the magnitude of χ(r   ); if
                       it is zero, then n    is a root of χ(s). Conversely, if A σ (x) has no root on the imaginary axis, then χ(s) cannot
                       have a root whose real part is the fixed value of σ from which A σ   is constructed.

                       Algorithm
                       Given N(s), D(s), h, σ min ,   and σ max :
                       Step 1. Pick σ values σ 1 ,…,σ M  between σ min  and s max  such that σ min  = σ 1 , σ i  < s i+1 , and σ M  = σ max . For
                            each σ i  perform the following.


                       ©2002 CRC Press LLC
   829   830   831   832   833   834   835   836   837   838   839