Applicability of failure criteria and empirical relations of mechanical rock properties from outcrop analogue samples for wellbore stability analyses

Knowledge of failure criteria, Young's modulus and uniaxial and tensile strengths, are important to avoid borehole instabilities and adapt the drilling plan on rock mechanical conditions. By this means, a considerable reduction of the total drilling costs can be achieved. This is desirable to enlarge the profit margin of geothermal projects which is rather small compared with hydrocarbon projects. Because core material is rare we aim at predicting in situ rock properties from outcrop analogue samples which are easy and cheap to provide. The comparability of properties determined from analogue samples with samples from depths is analysed by performing conventional triaxial tests, uniaxial compressive strength tests and Brazilian tests of both quarry and equivalent core samples. Equivalent means that the quarry sample is of the same stratigraphic age and of comparable sedimentary facies and composition as the associated core sample. We determined the parameters uniaxial compressive strength (UCS), Young's modulus, and tensile strength for 35 rock samples from quarries and 14 equivalent core samples from the North German Basin. A subgroup of these samples, consisting of one volcanic rock sample, three sandstone and three carbonate samples, was used for triaxial tests. In all cases, comparability of core samples with quarry samples is evaluated using thin section analyses. For UCS versus Young's modulus and tensile strengths, linear- and non-linear regression analyses were performed. We repeat regression separately for clastic rock samples or carbonate rock samples only as well as for quarry samples or core samples only. Empirical relations have high statistical significance and properties of core samples lie within 90% prediction bands of developed regression functions of quarry samples. With triaxial tests we determined linearized Mohr-Coulomb failure criteria, expressed in both principal stresses and shear and normal stresses, for quarry samples. Comparison with samples from larger depths shows that it is possible to apply the obtained principal stress failure criteria on clastic and volcanic rocks, but less so for carbonates. Carbonate core samples have higher strengths and develop larger angles between fault normal and main principal stress than quarry samples. This considerably reduces the residuals between quarry failure criteria and core test results. Therefore, it is advised to use failure criteria, expressed in shear and normal stresses, for prediction of core sample failure conditions. We conclude that it is possible to apply failure criteria on samples from depth if the comparability, especially textural comparability and similar porosities, of chosen outcrop analogues samples is ensured. Applicability of empirical relations of UCS with Young's modulus and tensile strength to rocks at depths is expected. Presented results may help predict mechanical properties for in situ rocks, and thus develop suitable geomechanical models for the adaptation of the drilling strategy on rock mechanical conditions. The authors appreciate the support of 'Niedersächsisches Ministerium für Wissenschaft und Kultur' and 'Baker Hughes' within the gebo research project (http: www.gebo-nds.de).