Page 211 - Rock Mechanics For Underground Mining
P. 211

FINITE DIFFERENCE METHODS FOR CONTINUOUS ROCK























              Figure 6.11 (a) Schematic finite dif-  one in which the problem unknowns can be determined directly at each stage from the
              ference grid; (b) representative zones  difference equations, in a stepwise fashion, from known quantities. Some advantages
              involved in contour integrals for a
                                        of such an approach are that large matrices are not formed, reducing the demands on
              gridpoint (after Brady and Lorig,
                                        computer memory requirements, and that locally complex constitutive behaviour such
              1988).
                                        as strain softening need not result in numerical instability. An offsetting disadvantage
                                        is that the iterative solution procedure may sometimes consume an excessive amount
                                        of computer time in reaching an equilibrium solution.
                                          Considering the two-dimensional case for simplicity, a finite difference scheme is
                                        developed by dividing a body into a set of convenient, arbitrarily-shaped quadrilateral
                                        zones, as shown in Figure 6.11a. For each representative domain, difference equations
                                        are established based on the equations of motion and the constitutive equations of
                                        the rock. Lumping of part of the mass from adjacent zones at a gridpoint or node, as
                                        implied in Figure 6.11b, and a procedure for calculating the out-of-balance force at a
                                        gridpoint, provide the starting point for constructing and integrating the equations of
                                        motion.
                                          The Gauss Divergence Theorem is the basis of the method for determining the
                                        out-of-balance gridpoint force. In relating stresses and tractions, the theorem takes
                                        the form
                                                                      %
                                                                          ij n j dS  (i, j = 1, 2)    (6.48)
                                                      ∂  ij /∂x i = lim A→0
                                                                       S
                                        where
                                          x i = components of the position vector
                                            ij = components of the stress tensor
                                          A = area bounded by surface S
                                          dS = increment of arc length of the surface contour
                                          n j = unit outward normal to dS.
                                        A numerical approximation may then be made to the RHS of equation 6.48, involving
                                        summation of products of tractions and areas over the linear or planar sides of a
                                        polygon, to yield a resultant force on the gridpoint.
                                          The differential equation of motion is

                                                                 ∂ ˚ u i /∂t = ∂  ij /∂x i +  g i     (6.49)
                                        193
   206   207   208   209   210   211   212   213   214   215   216