Page 739 - The Mechatronics Handbook
P. 739

0066_Frame_C23  Page 47  Wednesday, January 9, 2002  1:56 PM









                       Relationship between State Equations and Transfer-Functions
                       State-Space to Transfer-Function
                       The input-to-output relationship of a dynamic system in the frequency-domain is represented by a
                       transfer-function, which can be obtained by taking the Laplace transform of (23.107) and (23.108) with
                       zero initial conditions as follows [8, Chapter 3, section 5]:
                                                    sX s() =  AX s() +  BU s(),                (23.123)
                                                    Ys() =  CX s() +  DU s(),                  (23.124)

                       where s is the Laplace variable. Solving (23.123) for X(s) and substituting into (23.124), the ratio of the
                       output Y(s) to input U(s) for a single-input single-output system (SISO) can be found as
                                                       Ys()
                                                               (
                                                Gs() =  ----------- =  CsI A) B +  D
                                                                       1
                                                                      –
                                                                  –
                                                       Us()
                                                       Ns()
                                                     =  -----------                            (23.125)
                                                       Ds()
                       where I is an n × n identity matrix. In Eq. (23.125), N(s) and D(s) are referred to as the numerator and
                       denominator polynomial of G(s), respectively. 7
                         Analogous to the state-space equation, the boundedness of the output response y(t) to a bounded
                       input u(t) is characterized by the roots of the denominator polynomial D(s), i.e., the values of s for which
                       D(s) = 0. In particular, the output y(t) will be bounded for any bounded input, i.e., system is stable if
                                                                                8
                       the real parts of all the roots of D(s) are less than zero (strictly negative).  Alternatively, a convenient
                       method to determine stability without having to find the roots of D(s) explicitly is the Routh–Hurwitz
                       stability criterion [6, Chapter 6].
                       Example
                       With the state-space description of the mass-spring-damper system defined in Eqs. (23.112) and (23.113),
                       the transfer-function realization using Eq. (23.125) becomes

                                                                               –  1
                                         Ys()
                                  Gs() =  ----------- =  10 s  10  0      1       0   +  0 []
                                         Us()                –
                                                        01       ( –  k/m) – ( c/m)  b/m
                                               b/m
                                        =  ------------------------------------------          (23.126)
                                         s + ( c/m)s +  k/m
                                          2
                       The input to the system is the applied voltage V z (t) and the output is the displacement of the mass z(t).
                       Frequency-Response Using Transfer-Functions
                       Consider a linear single-input single-output (SISO) stable system with transfer-function description G(s).
                       When the system G(s) is excited by a sinusoidal input of the form


                                                       ut() =  Psin ( wt)                      (23.127)

                       with amplitude P and frequency ω, the output response (after the transients decay) will also be a sinusoid
                       of the form

                                                     yt() =  MPsin ( wt + f)                   (23.128)

                         7
                         The MATLAB command ss2tf can be used to convert a state-space realization to a transfer-function.
                         8
                         The MATLAB command roots (den) can be used to find the roots of den, where den is the coefficients of D(s).

                      ©2002 CRC Press LLC
   734   735   736   737   738   739   740   741   742   743   744