Page 226 - Acquisition and Processing of Marine Seismic Data
P. 226

4.4 FOURIER SERIES                             217






















           FIG. 4.4  Schematical illustration of convolution calculation of discrete time series x(t) and y(t) consisting of four discrete
           elements. τ is the convolution lag.


           cosine function summations, a periodic f(t) func-  while b n ¼ 0 for even functions in Fourier series
           tion is approached by                        summation (Table 4.1). Fourier series expansion is
                         ∞                              an approximation to the original periodic function
                   1    X
              ftðÞ ¼ a 0 +  ½ a n cos nω 0 tÞ + b n sin nω 0 tފ  and it allows us to express the periodic function
                                             ð
                                 ð
                   2
                        n¼1                             with several (in theory, infinite number of )
                                                  (4.6)  summed amplitudes of sine and cosine functions
                                                        with different frequency and phase characteristics.
           The basis of Fourier series expansion is to find  Although it is practically unfeasible, a perfect
           out a 0 , a n , and b n coefficients and substitute them  approximation to f(t)functionisobtainedifthe
           into Eq. (4.6). These three coefficients are  summationisdoneoverinfinitenumberofnvalues.
           obtained by following integral equations
                                                           For the Fourier series analysis, f(t) function
                     T=2                                should satisfy some specific conditions known
                  2  Z                                  as Dirichlet conditions:
              a 0 ¼     ftðÞdt
                  T
                     T=2                                 (i) f(t) function must be periodic and integrable
                     T=2                                    over a period, e.g., the power of the function
                  2  Z
              a n ¼     ftðÞcos nω 0 tÞdt n ¼ 0, 1, 2, …Þ   is limited over one period.
                              ð
                                      ð
                  T                                     (ii) f(t) function must have a finite number of
                     T=2                                    discontinuities and a finite number of
                     T=2
                  2  Z                                      extrema (finite number of maxima and
              b n ¼     ftðÞsin nω 0 tÞdt n ¼ 0, 1, 2, …Þ   minima) in a given time interval.
                                      ð
                              ð
                  T
                     T=2
                                                           As an example, let’s express a box-car (a
                                                  (4.7)
                                                        square wave) function as an infinite series sum-
           where a 0 is the mean value of f(t) function and is  mation of sine and cosine functions using Fou-
           known as DC component for n ¼ 0, which deter-  rier series expansion. Mathematical expression
           mines how much the summed harmonics are      of a box-car defined in the ( π, π) interval is
           shifted along the y axis. For odd functions, a n ¼ 0,  given by
   221   222   223   224   225   226   227   228   229   230   231