Page 80 - Fluid-Structure Interactions Slender Structure and Axial Flow (Volume 1)
P. 80

PIPES CONVEYING FLUID: LINEAR DYNAMICS  I               63

               similarly for the pipe problem, as U is increased. Hence, it is obvious that the frequen-
               cies  of  the  system must  decrease  with  increasing  U. At  u,,  the  lowest  (fundamental)
               frequency vanishes as the stiffness in that mode vanishes. In the linear sense, the original
               straight configuration becomes unstable,  and  all  adjacent deformed states in  that  mode
               become possible equilibria. In  the  nonlinear sense, a pitchfork bifurcation takes place,
               the original equilibrium is unstable and two stable equilibrium states, one on either side,
               emerge - defined by  the nonlinear forces acting on  the system, as will be demonstrated
               in Chapter 5.
                 However, the analogy of  the pipe with  supported ends to  the column with  the  same
               boundary conditions should not  be carried too far, because the latter problem is purely
               conservative, while the former is gyroscopic conservative. As will be shown later, despite
               the fact that the gyroscopic (Coriolis) forces do no work in the course of free oscillations,
               they do exert important influence on the overall dynamical behaviour.’
                 Finally, it should be mentioned that, according to linear theory, there should be no differ-
               ence in the dynamics of  systems (a) and (b) of Figure 3.1. In physical terms, however, it
               is obvious that buckling implies lateral deflection of the pipe. In system (b), once u 2 u,,
               the pipe may develop large static deflection since it is axially unrestrained. In system (a),
               on the other hand, where axial sliding of  the lower end is prevented, lateral deflection is
               associated with axial extension of  the pipe; this implies stretching and hence the gener-
               ation of  a deflection-related axial tension, a nonlinear effect. In practice, this means that
               the zero-frequency state is never achieved, as will be discussed further in Section 3.4.


               3.2.2  Cantilevered pipes

               As will be shown, a cantilevered pipe conveying fluid is a nonconservative system, which,
               for sufficiently high  flow velocity, loses stability by flutter  of  the single-mode type, i.e.
               via a Hopf bihrcation - see also Section 3.2.3.
                 The stability of cantilevered pipes conveying fluid [see Figure 3.l(c)] was first studied
               by  Bourrihres (1939),  who  examined the  problem  of  general motions  of  an  infinitely
               flexible and inextensible string, and the special case where the string is circulating (travel-
               ling) between two fixed supports; he  then tackled the problem of  one such string within
               another, which could have flexural rigidity - this of course being equivalent to the case of
               a pipe conveying fluid. He obtained the general nonlinear equations of motion, but did not
               develop them fully. Then, he linearized them and proceeded to study such diverse aspects
               as the difference between spontaneous and perturbation-induced instabilities (cf. Gregory
               & Paidoussis  1966b), and  the  wave  propagation  characteristics; he  also  attempted  to
               predict the period of self-excited motions, and studied several other aspects of the problem,
               as well as conducting experiments. On the other hand, he could not calculate the critical
               flow velocity, which, unlike the case of  a pipe with supported ends, requires the use of
               computersi - of  course, then unavailable. Bourrikes’  was  a truly admirable effort, and
               it is a pity that it was lost to posterity, until recently (Section 3.1). His work did not have

                 +In this respect, as civil servants the world over discovered long ago (and as viewers of BBC’s Yes Minisrer
               have witnessed to their delight), it is not necessary to do actual work in order to exert influence; see also Lynn
               & Jay (1989).
                 *Although Padoussis (1963). in order to check computer calculations - computers then  being a relatively
               new device - did do a hand calculation, thereby demonstrating its feasibility.
   75   76   77   78   79   80   81   82   83   84   85