Page 385 - Fluid-Structure Interactions Slender Structure and Axial Flow (Volume 1)
P. 385

PIPES CONVEYING FLUID:  NONLINEAR  AND CHAOTIC DYNAMICS        361


                           2.0


                           1 .5


                           I .0



                           0.5


                           0.0
                                                      8.027
                          -0.5   . . * * f . . . . I . *  .f'*.'"""".
                             7.90   7.95    8.00    8.05    8.10    8.15   8.20
                                                Fluid velocity, u

              Figure 5.36  The largest Lyapunov exponent of  the system of  Figure 5.34 versus u (Paidoussis &
                                              Moon  1988).

              significant result: that the correlation dimension in the chaotic regime is d, = 3.20. Here,
              it is recalled that noninteger values for dimension are perfectly normal for, and indicative
              of, chaotic systems.
                As shown by MaiiC (198 l), if d, is the correlation dimension of the system, the minimum
              number of  state variables, M, required for modelling the system is given by  d,  5 M  I
              2d, + 2, where M  corresponds to the dimension of  the system, i.e. to twice the number
              of degrees of  freedom N; therefore, this may be written as
                                          d, 5 2N  5 2d, + 2.                     (5.136)

              Both  M  and  N  have  to  be  integers  and  M  must  be  even  for  an  autonomous system.
              Therefore,  for  d, = 3.20, the  important result  is  obtained that  the  number of  degrees
              of  freedom required to capture the essential dynamics is 2 5 N  5 4 or 5, depending on
              exactly how  the  inequality is  interpreted. Hence, it is  not  surprising that  the  analytical
              results of  Figures 5.33-5.36  obtained with N  = 2 are in  qualitative agreement with the
              experimental observations. On the other hand, it is clear that, to capture the quantitative
              aspects of the dynamics adequately, an N  = 4 or N  = 5 model may be necessary.
                The first attempt to improve the analytical model aimed at (i) studying the dynamics
              of  the system for N  > 2 and (ii) conducting calculations with a more realistic model of
              impacting with  the constraining bars, by using their true stiffness and location - rather
              than  the  strained ('relaxed')  values used  in  the  foregoing, resorted to  solely to  obtain
              convergent solutions. The  measured  stiffness of  the  constraining bars  is  very  close to
              the trilinear model, as shown in Figure 5.38. With this model, the cubic spring term in
              equation (5.131) is replaced by

                       f(q>s($ - (b),   f(q) = K{q - $(lq + qbl  - Ir  - qbl)}.   (5.137)
   380   381   382   383   384   385   386   387   388   389   390