Finite element analysis of two overcoming techniques for in situ stress measurement

Abstract

The C.S.I.R. “doorstopper” technique and the U.S.B.M. borehole deformation technique are analyzed systematically using the finite element method. Displacement discontinuity introduced by the overcoring process is simulated by the “split node” technique. The axisymmetric (equal biaxial) loading problem and non-axisymmetric (pure shear) loading problem are analyzed separately, the general case being a superposition of the two. For the “doorstopper” method, stress concentration factors are calculated for various borehole lengths as well as different Poisson's ratios. Computation shows that the previously calculated stress concentration factors are applicable only if the borehole length is longer than three times the radius. The local stress relief behaviour is influenced principally by the stress concentration factors. A comparison with data indicates that measurements of the stress relief behaviour can potentially be used to assess the quality of field data. Previous misconceptions concerning interpretation of the U.S.B.M. borehole deformation data have been clarified by the finite element calculation. The results show that the relation between tectonic stresses and the permanent radial displacement at the pilot hole resulting from overcoring is given by the “plane stress” solution for measurements made right at the base of a borehole. For measurements made deep inside the pilot hole, Leeman's [29] “three-dimensional” solution instead of the “plane strain” solution is applicable. Stress concentration factors are defined and calculated for various measuring depth, borehole length and Poisson's ratio. These factors decrease with measuring depth. At the base of a borehole, the concentration factors are almost identical to those for the “doorstopper” configuration. The stress concentration factors are equal to one for measurements deep inside the pilot hole. It is therefore suggested that future U.S.B.M. borehole deformation measurements ought to be made at locations corresponding to either of the two limiting cases for which the calculations of in situ stresses from data are relatively unambiguous.