Scale or size effects of the uniaxial strength response in rock samples have been studied in detail in the past, so a good number of studies on different rocks are available. However, analyses on triaxial strength scale effects in rocks are scarce and they seldom address failure criteria (i.e. Hoek-Brown) evolution with specimen size. This obvious lack of data can be attributed to the difficulties of having available Hoek’s cell of different sizes. With the aim of filling this void, the authors have carried out sets of around 25 stress-strain triaxial tests on intact 30, 38, 54 and 84 mm diameter granite specimens, with various confinements (0.2 to 15 MPa), so reliable estimates of Hoek-Brown strength were obtained for every scale. We compare results with previous studies on UCS scale effects, showing a good correlation. Results suggest the studied granite undergoes a reverse size effect in terms of strength at low confinements. Indeed, the UCS increases as sample diameter increases up to around 50 mm, but decreases thereafter. However, results obtained put forward that this strength variation with scale tends to be mitigated for higher confinements where the scale effect may not be clearly recognised. So increased confinement can be associated with a decreased scale dependency component of strength.
The mechanical behavior of rocks at micro and macroscopic scale is complex and inhomogeneous, something the authors attribute to geologic conditions, tectonic and weathering effects as well as to hydro-thermo-chemical processes. This is why it is not surprising that the mechanical response of rock specimens depends on the volume of material being loaded or strained. That is, rock strength shows significant scale effects. These effects are not well understood, yet they arise pervasively in rock mechanics and engineering practice in the determination of rock, joint and rock-mass properties.
Pinto da Cunha [1] emphasized the 3 different causes for mechanical parameter variation in rocks and rock masses linked to scale including: a) effects associated with jointing, typical of rock masses, b) shape effects caused by changes in system geometry and typically in specimen slenderness (a height to diameter ratio of at least 2:1 is recommended as per ISRM [2]) and c) intact rock size effects occurring unrelated to effects associated with geometry or jointing, which are the focus of this study.