Page 112 - Modern Control Systems
P. 112

86               Chapter 2  Mathematical  Models of Systems



                            *.o





           FIGURE 2.30      «:0
           Signal-flow graph
           of two algebraic
           equations.



                            and
                                                (1 -  a u)r 2  +  a 2lri
                                        * 2  =                         - ^ - r 2  +  —r h      (2.94)
                                             (1  -  fln)(l  -  fl 22)  -  «i2«2i
                               The  denominator  of the  solution  is the  determinant  A of  the  set  of  equations
                            and is rewritten as
                               A =  (1  -  fl n)(l -  fl 22) -  «12«21 =  1 ~  «11 -  «22 +  «11«22 -  «12«21-  (2-95)
                            In  this case, the denominator  is equal to  1 minus each  self-loop  a n,  a 22, and  fli2«2i>
                                                                                           an
                            plus the product  of the two nontouching loops a\\  and a 22. The loops  #22  d  «21 «12
                            are touching, as are an  and  fl2i«i2-
                               The numerator for X\ with the input r\  is 1  times 1 —  a 22, which is the value of A
                            excluding terms that touch the path 1 from r\ to x\. Therefore the numerator from r 2
                            to Xi is simply a 12 because the path through a\ 2 touches all the loops. The numerator
                            for x 2 is symmetrical to that  of X\,
                               In  general,  the  linear  dependence  7^- between  the  independent  variable  x t
                            (often  called  the input  variable)  and  a dependent  variable  Xj is given  by Mason's
                            signal-flow  gain formula [11,12],
                                                             2^* -/;*

                                                        T  =                                  (2.96)
                                Pijk  =  gain  of kth  path from  variable x t  to variable Xj,
                                 A =  determinant  of the graph,
                                Aijk =  cofactor  of the path  P ijk,

                            and the summation is taken over all possible k paths from  x t  to Xj. The path gain or
                           transmittance P^  is defined  as the product  of the gains of the branches of the path,
                           traversed  in the direction  of the arrows with no node encountered more than once.
                           The cofactor  A^  is the determinant with the loops touching the kth path  removed.
                           The determinant  A  is
                                               L
                                     A = 1 -  2 «  +    £j   LnLffj      ^j  L-'n'-'m'-'p  '  (2.97)
                                             «=1        n,m            n, m, p
                                                     nontouching     nontouching
                           where L q equals the value of the qth. loop transmittance. Therefore  the rule for eval-
                           uating A  in terms  of loops L l5 L 2, L 3 ,..., L N  is
   107   108   109   110   111   112   113   114   115   116   117