Application of the Hoek-Brown failure criterion to New Zealand hard greywacke rocks leads to unrealistically high predictions of rock mass strength for better quality rock masses and lower predictions for poorer quality rock masses. To further investigate linkages between rock masses in the field and failure criterion inputs such as the Geological Strength Index (GSI), a descriptive engineering geological classification for unweathered greywacke with five classes has been implemented. The results of laboratory triaxial testing of lower quality (Class IV) rocks with extremely-closely spaced defects (c¢=0 kPa, f¢=48°) and assessment of in situ direct shear tests on better quality (Class II) rocks that give similar strengths show further calibration of strength data against the failure criterion is required.
These Upper Paleozoic to Mesozoic-age rocks are widely spread throughout New Zealand (Figure 1) and are an important source of roading and concrete aggregate as well as being the basement rocks in many of the country's engineering projects. Greywacke rocks are commonly closely jointed as a result of their complex tectonic and geological history. The Hoek-Brown failure criterion (Hoek & Brown, 1997) is an often used method of assessing the strength of closely jointed rock masses, but experience in New Zealand has shown that predictions from this criterion cannot be confidently applied to greywacke rocks. The requirement for a workable failure criterion has led to this research project into the engineering properties of unweathered greywacke. In the first stage of the project, three dam or quarry study sites (Aviemore, Belmont, Taotaoroa – Figure 1) were engineering-geologically mapped and laboratory testing was carried out to define the Hoek-Brown constants for intact rock (Read et al., 1999). This present paper describes further assessment of the failure criterion based on mapping, sampling and laboratory testing largely from Aviemore.