Statistical Stress Model Calibration

The estimation of orientation and magnitude of crustal stresses is crucial for the design phase of technological and safe underground usage. As usually only a few in-situ data are available, stress prediction is challenging. Geomechanical-numerical modelling is the only tool, which allows stress prediction that takes material properties and inhomogeneities into account. The model calibration is essential to find the best-fitting stress results. The presented workflow allows the statistically proved calibration to determine the best-fit model. Furthermore, statistic tests speed up the calibration process. The order in which model-independent data are used for model calibration pay attention to the interrelation between the stress components. Therefore, data of vertical stress are tested first to optimize material properties. Second, the orientation of the maximum horizontal stress is used to optimize orientation of applied boundary conditions. Finally, the magnitudes of minimum and maximum horizontal stress are varied to find the optimal strain, applied by the boundary conditions.