Page 715 - Automotive Engineering Powertrain Chassis System and Vehicle Body
        P. 715
     CHAP TER 2 1. 1       Interior noise: Assessment and control
               Differential equations                               The same rules apply to the well-known, second-order
                                                                  differential equation characterising the motion of a mass
               The input–output relations for networks that store  on a spring with a viscous damper
               energy (dynamic systems) are given by differential   m€ x þ c_ x þ kx ¼ FðtÞ              (A21.1.8)
               equations.
                 Consider a linear, continuous-time system where:   If two initial conditions are known, namely x(0)
                 The output is represented by y(t);               and € xð0Þ, the value of x at some instant t later may be
                 The input is represented by x(t).                found.
               The two variables will be related by a differential equa-
               tion of the form:
                                                                  Appendix 21.1B: The convolution
                    n        n 1
                   d y      d   y         dy
                 a n  þ a n 1    þ / þ a 1  þ a 0 y               integral
                   dt n     dt n 1        dt
                        m
                       d x
                   ¼ b m   þ / þ b 0 x                (A21.1.1)   Some aperiodic signals have unique properties and are
                       dt m
                                                                  known as singularity functions because they are either
                 This equation is called a ‘linear differential equation of  discontinuous or have discontinuous derivatives (Sinha,
               order n’ and for most practical cases n P m.       1991). The simplest of these is the unit step function (as
                                           d
                 It is convenient to replace  dt  by the operator ‘p’  illustrated in Fig. B21.1-1), given the symbol g(t)
               resulting in the equation                            The unit impulse function or delta function d(t)is
                                                                  defined as the function which after integration yields the
                    n
                 ða n p þ a n 1 p n 1  þ / þ a 1 p þ a 0 ÞyðtÞ    unity step function, so (Sinha, 1991)
                         m
                   ¼ðb m p þ / þ b 1 p þ b 0 ÞxðtÞ    (A21.1.2)            ð t
                                                                    gðtÞ¼      dðsÞds                     (B21.1.1)
               which may be written compactly as                             N
                 DðpÞyðtÞ¼ NðpÞxðtÞ                   (A21.1.3)   Alternatively,
                                                                          gðtÞ
               where D( p) and N( p) are polynomials in the         dðtÞ¼                                 (B21.1.2)
               operator ‘p’                                                dt
                                                                    The impulse function must satisfy:
                           n
                 DðpÞ¼ a n p þ a n 1 p n 1  þ / þ a 1 p þ a 0
                                                      (A21.1.4)     dðtÞ¼ 0; for t not equal to zero      (B21.1.3)
                           m
                 NðpÞ¼ b m p þ b m 1 p m 1  þ / þ b 1 p þ b 0     and
                                                      (A21.1.5)     ð
                                                                     N
                                                                        dðtÞdt ¼ 1                        (B21.1.4)
                 The operator ‘p’ does not satisfy the commutative
                                                                      N
               property
                                                                    Therefore, the area under the impulse function is
                 pyðtÞsyðtÞp                          (A21.1.6)   unity and it occurs over an infinitesimal interval around
                                                                  t ¼ 0. So, as the period dt tends towards zero, the height
                 The system operator L( p) or transfer function is the  of the impulse function approaches infinity.
               ratio of the two polynomials D( p) and N( p)
                                                                    Also,
                        NðpÞ
                 LðpÞ¼                                (A21.1.7)     ddðtÞ
                        DðpÞ                                             ¼ N    at  t ¼ 0 and is zero elsewhere:
                                                                     dt
                 Any dynamic system described by a differential
               equation of order n can be solved uniquely only if at least
               n initial or boundary conditions are known.                 x(t)
                 As an example, consider that the unique solution to
                                                                           1
               the differential equation characterising the input–                              x(t) = 0, t<0
               output relationship of an electrical circuit can be                              x(t) = 1, t>0
               obtained only if the initial values of the voltages across
               each capacitor and the current through each inductance                        t
               are known.                                         Fig. B21.1-1 The unit step function.
                    726
     	
