Page 362 - Numerical Analysis and Modelling in Geomechanics
P. 362

RESERVOIR COMPACTION, SUBSIDENCE AND WELL DAMAGE 343





















            Figure 11.5 The flow potential surface.

                          Constitutive model input data requirements
            As  indicated  above,  a  large  number  of  material  parameters  can  be  required  to
            specify  the  complete  behavior  of  a  geomaterial,  which  requires  numerous
            complicated  and  often  expensive  laboratory  tests.  A  listing  of  the  material
            parameters  required  for  a  coupled  geomechanical  analysis  of  reservoir
            compaction using the ABAQUS program modified Drucker-Prager/Cap model
                              83
            is shown in Table 11.1.  Weakly cemented and unconsolidated clastic rocks and
            soils exhibit significant elastic nonlinearity at low stress, due to the nonlinear grain-
            to-grain contact deformation or closing and sliding of micro-cracks. A nonlinear,
            logarithmic  bulk  modulus,  K,  may  be  used  to  simulate  this  behavior.  In  most
            cases,  this  level  of  detail  is  not  necessary,  and  a  linear  elastic  modulus  would
            provide  acceptable  accuracy.  In  some  cases,  particularly  for  highly
            unconsolidated rocks and sandy soils in the vadoze zone, the pore fluid may be
            compressible  and  the  grains  may  be  considered  compressible.  In  the
            development  of  the  finite  element  equations  given  in  the  previous  section,  the
            assumption  was  made  that  the  grains  were  incompressible  and  the  pores  were
            fully saturated. Most commonly, for reservoir scale compaction simulations, the
            elastic behavior is modeled by providing the Young’s modulus and the Poisson’s
            ratio  as  determined  from  uniaxial  or  triaxial  compressive  stress  tests.  For
            unconsolidated  materials,  uniaxial  compressive  tests  may  be  difficult,  if  not
            impossible.
              The post-yield or post-failure behavior of rock is difficult to determine easily
            and inexpensively. The shear failure surface may in fact be nonlinear. Typically,
            at  least  three  triaxial  compression  tests  are  required  to  establish  the  Drucker-
            Prager  shear  failure  surface  and  its  hardening  behavior.  A  single  hydrostatic
            compression  test  involving  loading  and  unloading  multiple  times  could  be
            conducted to determine the compaction evolution curve for the Cap surface and
   357   358   359   360   361   362   363   364   365   366   367