Abstract
Four failure criteria, a simplified Mohr–Coulomb, a Mohr–Coulomb, a von Mises and a Drucker–Prager model are considered for borehole failure and sand onset predictions. The von Mises is a special case of the Drucker–Prager. The resulting analytical expressions are suitable for implementation in the analytical software for sand onset and sand mass analyses. The models are calibrated and validated against experimental data for hollow cylinder and hollow prism tests on Red Wildmoor sandstone. The analysis shows that a Drucker–Prager model is necessary to capture the experimentally observed effect of axial stress on hole failure. The simplified Mohr–Coulomb model shows no such effect, while the Mohr–Coulomb model shows an effect only at high values of axial stress. The Drucker–Prager model can match satisfactorily the experimental results. It requires, however, the calibration of an additional material parameter which can be done on anisotropic loading hollow prism tests. The models include also a lateral stress anisotropy parameter and a field calibration factor to allow for better approximation of available field or laboratory data.