Most conventional rock mechanics tests have their origin in civil engineering, where design rules necessarily dictate a conservative approach. For the petroleum industry applications however, partial rock failure may be acceptable in order to optimize design. If such approaches are to be adopted in a manner other than purely empirically then the method of stability analysis must fully account for the failure behaviour of intact rock; i.e. the transition from peak to residual strength including the change in friction, dilatancy and cohesion characteristics and not least, the role of the intermediate principal stress. A number of polyaxial tests have been carried out on Saltwash North and Castlegate outcrop sandstone, and compared to results from conventional triaxial tests. The test results have been back analyzed using the ISAMGEO FE code. The results of the study confirm that the intermediate principal stress does have an impact on rock failure and that this is not adequately captured by the standard Mohr-Coulomb criterion. In response a modified Mohr-Coulomb criterion is established. Furthermore, using a simple example of openhole stability, the study demonstrates that the results of numerical modeling are sensitive to the method adopted during model calibration in addition to accounting for the intermediate principal stress and post-peak strength strain softening. These effects will have a critical impact on engineering design if this is to include partial rock failure.
Most conventional rock mechanics tests have their origin in civil engineering where design rules necessarily must be conservative. As such, both tests and geotechnical design have primarily focused on pre-failure behaviour where the rock is largely intact and elastic, and the determination of the peak strength. In contrast, for some oil reservoir applications partial rock failure may be acceptable in order to optimize design. Under some drilling applications partial failure of the wellbore wall (breakout development) may be acceptable or even desirable in order to reduce mud weight provided there is adequate hole cleaning. Alternatively in the design of multilateral well junctions significant cost reductions can be achieved if the junctions are completed without cemented casings . Under these circumstances, a viable stability analysis must properly account for the post-peak strength behaviour, dilatancy and the effect of the intermediate principal stress. To meet this challenge, new and more advanced laboratory tests are required to provide the necessary information and numerical models must become more refined.
A number of polyaxial tests have been carried out on Saltwash North and Castlegate outcrop sandstone, and compared to results from conventional triaxial and uniaxial rock mechanic tests. The polyaxial test parameters were determined through pre-test numerical modeling where the intention was to design stress paths that simulated near-wellbore stress conditions. Subsequently the results from both the conventional and polyaxial tests were used in numerical analysis for the purposes of evaluating model refinement and constitutive model formulations where particular emphasis was placed on matching peak strength and post-peak stress-strain behaviour.
2. LABORATORY TESTING
A set of conventional and polyaxial rock mechanic tests were performed on the two outcrop sands used in the study, Castlegate and Saltwash North.