Previously borehole breakout and drilling induced tensile fracture data have been analysed to determine the state of stress in the Cooper-Eromanga Basin of South Australia. It has been shown there is considerable variability about a mean maximum principal stress direction of 101° N, with a transition in the stress state varying across the region from a reverse to strike-slip and normal regime. At great depths (>2.5 km) the two most difficult parameters to constrain when undertaking numerical stress modelling are (1) fault stiffness, which influences how a fault behaves under a particular regional stress regime; and (2) the magnitude of the minimum (Shmin) and maximum (SHmax) horizontal principal stresses. It has been shown that variability in breakout data can be actively used to calibrate numerical models. The breakout data, are usually collapsed to a single data point for any one well and graded for quality. This grading is highly dependent on the variability of the maximum horizontal stress direction, with the resultant orientation for any one well being downgraded as the variability in the data set increases. A 2D distinct element model, comprising 107 faults interpreted from seismic data, was built at basement depths for the province. Correlations were made between the modelled and analytically derived fault displacements and the modelled stress rotations with well data. The results have shown good correlation with breakouts where faults are associated, which has provided information on the relative strength of faults across the region. Importantly the model can provide information relating to the risk of fault reactivation in the region and, where well control is poor, can provide confidence in predicting stress variability across the region. Also the full model for this region is applicable to future petroleum production analysis, as regions of known high mean stress surrounding areas of low mean stress create pressure gradient potentials and may indicate regions of hydrocarbon pooling. The method would be applicable to other regions.
Stress modelling is being developed as a useful tool in the field of petroleum geomechanics. The different methods for numerically modelling the effect of heterogeneities on stress include finite element method (FEM), boundary element method (BEM), finite difference method (FDM), combinations of FEM/FDM and distinct or discrete element method (DEM). These methods and their general applicability have been reviewed by Jing and Hudson , and Jing . The computational techniques have been written into various codes or software and are routinely used in the field of mining and geotechnical engineering. In contrast to the petroleum industry, the depths of investigation are often shallow and many of the parameters required for accurate modelling can be directly measured.
Typically, the first stage of any geomechanical study in the petroleum industry involves determining the constraints on, or determination of the regional stress regime. Inherent uncertainty in this process is due to the following factors:
? depth of investigation, usually in excess of 1 km
? lack of data from which the magnitude of minimum horizontal stress (Shmin) may be
measured, for example extended leak off tests and/or mini-frac data