: The prediction of in situ failure conditions has been a critical issue in rock engineering for decades. Various approaches have been developed and used to estimate the strength of rock media around openings at different scales. In this paper, the authors present a novel approach based on a statistical analysis of laboratory data, to evaluate the large scale strength sL of a unit block of rock. The strength of the actual rock mass is then described by applying a phenomenological expression that makes use of the RMR geomechanical classification and of sL. The procedure is used with the MSDPu criterion to define the failure surface of the rock mass. After recalling the MSDPu formulation and introducing the proposed scaling functions, a sample application is briefly presented and discussed.
Estimating the in situ rock mass failure conditions is still problematic due, in part, to the difficulty in obtaining appropriate large scale strength values. Representation of rock mass behavior is a complex issue, that involves various and interrelated mechanical phenomena including the influence of scale on the response of "intact" rock and the effect of structural defects (joint sets and faults) which are not present in small scale specimens tested in the laboratory. The large scale failure strength of rock mass nevertheless depends on these two factors, and any scaling function used, to go from laboratory to in situ, should take them into account.