Page 168 - Fluid-Structure Interactions Slender Structure and Axial Flow (Volume 1)
P. 168

150               SLENDER STRUCTURES AND AXIAL FLOW

                    along the length of the pipe; a corresponding term appears in the shear boundary condition
                    at the clamped end. The new term in the equation of motion opposes the centrifugal term,
                    U2(a2w/ax2), and this generates a strong stabilizing effect. Adding foundation damping
                    in this case is destabilizing, although these results were confined to B =   only, and
                    hence should not be considered as general.
                      All the foregoing apply to uniform foundations. Unusual behaviour may be expected
                    for nonuniform ones, however, in view of Hauger & Vetter’s (1976) results for the system
                    with the follower load (pipe with j?  = 0) and k  = k(t). Thus, if k(t) is zero at  = 0 and
                     1 and  triangularly distributed in-between, with  a maximum at 4 = ern of  k,,,   then the
                    effect is  destabilizing for all k,,,.  The opposite is true if  k  = 0 at 6 = tm, and kmaX at
                    [=Oand  1.


                    3.5.8  Effects of tension and refined fluid mechanics modelling

                    The  system  shown in  Figure 3.57(a)  was  studied  by  Guran  & Plaut  (1994), in  which
                    the compression P  is conservative, i.e. it is constant and remains along the undeformed
                    axis  of  the  pipe.  The  equation of  motion  is  equation (3.98) with  Tj = 0 and T = -P,
                    or r = -9 in  dimensionless form,  where 9 = PL2/EI. The  boundary  conditions, in
                    addition to ~(0, t) = 0 and (a2q/at2)1+,  = 0 are




                    where K*  = CL/EI. Clearly, in the limit of  K*  = 00  the pipe becomes cantilevered.
                      Typical results for  K*  = co are shown in  Figure 3.57(b), and  it is clear that  they  are
                    similar to those of  Figure 3.33. Indeed, the physical systems are similar: in the case of
                    Figure 3.33 the pipe is subjected to conservative gravity-induced distributed compression,
                    but in  this case to conservative uniform compression along the length.
                      Results for  K*  # 00  are  similar. The  influence of  K*  on  stability is  given in  graph-
                     ical form, for both divergence and  flutter, in  Guran  & Plaut (1994). The condition for
                     divergence is
                                   u2 + 9 cos v - (Vp)/K*) sin u = 0,  v = v’2TiF.      (3.111)

                       As  suggested  by  Figure 3.57(b)  and  as  may  be  verified  numerically  with
                    equation (3.1 I l), if the system is subjected to conservative fension it cannot lose stability
                    by divergence, but only by flutter.
                       The  effect  of  the  small  tension  induced  by  the  presence  of  a  terminal nozzle  on  a
                    cantilevered pipe - refer to the discussion associated with equations (3.40)-(3.42)  - was
                     taken into account by  Bishop & Fawzy (1976), who found
                                                TL = iMU2(~j - 1) 2                     (3.1 12)

                     by taking a force balance across the nozzle, where a, = A/Aj is the ratio of pipe flow area
                     to nozzle terminal flow area. This tension  was  neglected in  the  theoretical calculations
                     to  which  the  experiments  of  Gregory  & Paldoussis  (1966b) were  compared. As  seen
                     in (3.112) this  is  not  necessarily negligible for  crj  substantially different from  1; it  is
                    properly taken into account in the comparison in Figure 3.53. A  more refined treatment,
   163   164   165   166   167   168   169   170   171   172   173