Page 128 - Mathematical Techniques of Fractional Order Systems
P. 128

116  Mathematical Techniques of Fractional Order Systems


               Proof: First of all, it is possible to write (4.66) in the form

                       k  ð 1N
                      X        b k;i
                                     τ k21
                                            i
                                                     ðiÞ
                p k tðÞ 5     k 2 1Þ!  e τ  ln t 2 τÞ 2 Γ ðÞ Ht 2 τÞdτ  ð4:67Þ
                                                       1
                                            ð
                                                           ð
                      i51  0  ð
            in which H :ðÞ denotes the Heaviside unit step function. Taking the Laplace
            transform of (4.67) and using the time shift property of the Laplace trans-
            form yields
                      k  ð 1N
                                    τ k21 2sτ
                     X        b k;i
                                                   i
              P k sðÞ 5      k 2 1Þ! e τ  e  L t-s  ln tðÞ 2 Γ ðÞ HtðÞ dτ
                                                         ðiÞ
                                                           1
                     i51  0  ð
                                                                      ð4:68Þ
                                     k
                  1N    1    τ k21 2sτ  X        i
               5            e τ  e     b k;i L t-s ln tðÞ 2 Γ ðÞ=s dτ
                 Ð
                                                          1
                                                        ðiÞ
                  0   k 2 1Þ!
                     ð
                                     i51
               Using Lemma 6 gives
                                                             τ k21
                      ð 1N   1               k         k  ð 1N  e τ
                                  τ k21 2sτ
               P k sðÞ 5         e τ  e   ð lnsÞ  dτ 5  ð lnsÞ     e 2sτ dτ
                       0  ð k 2 1Þ!         s        s   0  ð k 2 1Þ!
                                                                      ð4:69Þ
               Note that the integral in the right side of (4.69) is actually the Laplace
                         n     o
                          e τ
                           τ k21
            transform L τ-s  k 2 1Þ!  . Therefore, using Lemma 2 gives
                          ð
                                                 k
                                             ð lnsÞ
                                     P k sðÞ 5                        ð4:70Þ
                                            ss21Þ k
                                             ð
               On the other hand, note that taking the Laplace transform of p k tðÞ by
            directly using its definition (i.e., (4.52)) would result
                       Ð    t  τ    k  τ
                  L t-s  e E ðÞdτ 5
                            1
                        0
                                                             k
                  1         k  	      1          	  k      ln sðÞ     ð4:71Þ
                            t
                    L t-s E ðÞ     5    L t-s E 1 tðÞ   5
                          1
                  s           js-s21  s            js-s21  ss21Þ k
                                                           ð
               By comparing (4.70) and (4.71), the proof of this Lemma is
            completed.
               Prior to proceeding, it is worth mentioning that calculation of convolu-
            tions involving exponential integral functions are of interest in some papers
            (Fisher and Al-Sirehy, 2015; Geller and Ng, 1969). This problem is treated
            in the following Remark as a side result of Lemma 7, which concerns the
            k-fold convolution of exponential integral function with itself. This gener-
            alizes the work done in Geller and Ng (1969).
   123   124   125   126   127   128   129   130   131   132   133