The Meaning of Average Compressive and Tensile Strength for Hoek-Brown mi Constant Determination

One of the most widely used failure criteria for rocks in the world is the Hoek-Brown failure criterion. For its use, the mi empirical parameter for a specific rock type is needed. The triaxial compression test is recommended for its determination; however, the full stress path for every rock comprises confined tension as well. This affects the course of the Hoek-Brown envelope, which is non-linear and starts at uniaxial tension. Fifty-one series of tests were carried out for three rock types: sandstone, claystone and limestone, to show the difference between the results of the mi determination, using two different approaches – so-called linear and non-linear. Moreover, the consistency between the developed simplified methods of constant determination and mi were checked. These comprised the UCS-based method, R-index method, TS-based method and advanced regression functions of compressive and tensile strength. The relationship between mi constant and the internal friction angle was checked as well. The analysis of the results showed that the consistency with the regression models developed by researchers depends on the chosen estimator. If it is derived from the triaxial test only, the results are closer to a linear determination of mi constant and have a good correlation with internal friction angle. If both tensile and compressive strength are used for its determination, the non-linear value correlates better with the advanced regression functions, but quite poor with the average compressive strength (R-index method) and tensile strength (TS-based method). Taking into account that every rock retained next to the geotechnical or mining object is not only compressed but also tensed, the non-linear mi interpretation seems to be more correct. The interlayers and discontinuities inside sedimentary rocks increase the scatter of lab results and reduce the accuracy of mi determination.