Page 181 - Numerical Analysis and Modelling in Geomechanics
P. 181

162 C.L.RAMSHAW AND A.R.SELBY
            with  the  beams  but  not  rotation.  This  model  could  not  compute  geometric
            attenuation  of  the  ground  waves,  so  the  strategy  used  was  to  input  reasonable
            impact or vibratory excitation around a pile space, and to compute pipe strains
            when the ppv’s at the ground surface directly above the pipe were 10 mm/s. This
            relates  to  Transco’s  field  practice.  Figure  5.25  shows  an  FE  mesh  for  a  10  m
            deep pipe and an impact hammer, and an FE/IE mesh for a 2 m deep pipe with
            vibrodriver.
              A number of cases were analysed, and some peak bending stresses induced in
            the pipe cross-section are summarised in Table 5.3.
              The stresses induced into the pipe by bending deformation of the cross-section
            are very small, and are negligible in comparison with the yield stress of the pipe
            steel  of  413  MPa  or  552  MPa  respectively.  There  is  strong  evidence  that  the
            stresses  induced  by  deformation  of  the  cross-section  are  negligible.  However,
            other  forms  of  distress  to  pipelines  may  be  caused  by  wave-generated
            mechanisms, such as joint pull-out or pipe-branch joint damage.
              The  deformed  shapes  of  the  pipe  cross-section  in  response  to  P,  S  and  R-
            waves were particularly interesting and not entirely expected. The passage of a P-
            wave  was  fairly  predictable  in  causing  ovalling  of  the  cross-section  with  the
            major axis normal to the direction of travel of the wave front, Figure 5.26. This
            deformation caused the lowest stress during passage of ground waves.
              The  response  of  the  cross-section  of  the  pipe  as  an  S-wave  passed  through
            showed a maximum deformation with a major axis at 45° to the direction of the
            wave  travel.  This  can  be  verified  by  consideration  of  the  effect  of  an  S-wave
            upon a square element, as shown in Figure 5.27.
              Finally,  deformation  of  the  pipe  section  near  to  ground  surface  caused  by  a
            Rayleigh  wave  is  even  more  complex,  as  can  be  anticipated,  because  of  the
            combination  of  vertical  and  horizontal  movements  of  a  soil  particle  in  a
            retrograde elliptic path, see Figure 5.28.

                                       Conclusions

            Impact  driving  of  a  pre-formed  pile  causes  significant  transient  ground
            vibrations, which attenuate non-uniformly with distance from the pile. Vibration,
            expressed  as  ppv’s,  is  a  function  of  hammer  energy,  hammer-pilehead  impact,
            and, to a lesser extent, of soil type.
              A  three-stage  computational  model  has  been  shown  to  be  capable  of  good
            quality  back-analysis  of  the  impact  and  of  the  definition  of  outgoing  ground
            vibrations. The method is less reliable in pre-estimating vibrations because of the
            variability  of  driving  parameters,  of  soil  properties  and  of  site  construction
            features.  The  selection  of  an  appropriate  dynamic  soil  modulus  has  been
            discussed.
              Pile installation and extraction by vibrodriving is highly effective in saturated
            fine  sands  and  silty  soils.  A  two-stage  computational  procedure  has  been
            developed  which  shows  close  agreement  with  site  records  of  ground  surface
   176   177   178   179   180   181   182   183   184   185   186