Page 464 - Fluid-Structure Interactions Slender Structure and Axial Flow (Volume 1)
P. 464

436                SLENDER STRUCTURES AND AXIAL FLOW

                    6.3.2 Analysis for extensible pipes

                    Because the static displacements appear explicitly in the equations of  motion, the static
                    equilibrium is determined first, and then the stability of  motions about this position.

                     (a) Determination of the static equilibrium
                    In order to discretize the equations governing the static deformations, the following vari-
                    ational statement is utilized


                                                                                        (6.100)
                                     j= 1

                    where Sqp and 6q;  are the variations in the steady-state displacements q:  and q:,  while
                    Afl(g:, q;)  and    qz) represent the  left-hand  sides of  equations (6.69)  and  (6.70),
                    respectively; n  and <,  are as in the foregoing.
                      The solutions for   and qi are sought in the form




                     where [Nle] and [N3e] are two matrices of  interpolation functions of  the space coordinate
                     <,  and   is the element in-plane displacement vector. It may be  shown that  a cubic
                     interpolation model for  qy  and linear interpolation for qi can guarantee convergence of
                     the  finite element  scheme. Thus,  one  can proceed in  the  same manner as for the  out-
                     of-plane motion in the inextensible case to obtain a matrix equation governing the static
                     equilibrium of an element as follows:t

                                                 [Kp]"(q;)"  = (Fpje,                   (6.102)

                     where
                             =
                        [qe ([Aol-')T  [{[M + @([1151 + [1l5lT> + 02[1*1)
                               + de([I81 - @([[I21  + [112JT)  + @2[111)
                               + (npo -I- z2)([191 4-  @([I221 - [In]) - 02[141)        (6.103)
                                 -
                               - hE2{[151 + [I101 + @([116l + [I141 - [1171) - @2[1i,l)] [&I-',
                                                         -
                        (FpJe =        [@(npo + z2){F1} - hZ2(@{F2) + {F3)) + (F4)] .

                     For the integrals [ZI], (Fl}, etc. and [A,,], see Appendix K.
                       In deriving equation (6.102) it has been assumed that the pressure in the external fluid
                     is constant, while the internal pressure varies linearly along the centreline. Thus


                                                 np = nplo - ZE2(,                      (6.104)

                       +It is recalled that, for an inextensible pipe, in-plane motion involves sixth-order derivatives in <,  and hence
                     the corresponding shape functions are not useful here. However, the shape functions for out-of-plune motions
                     of the inextensible pipe may be  used for this analysis, since the maximum orders of partial derivatives match;
                     hence the appearance of  [A,] in equations (6.103).
   459   460   461   462   463   464   465   466   467   468   469