Eurocode 7 requires the characteristic values of geotechnical parameters to be selected as cautious estimates of the values affecting the occurrence of the limit state. In this paper we re-analyse some historical strength data of Delabole slate and thus show some difficulties associated with characterisation of the strength of anisotropic rock. The strength data are grouped according to the orientation of the cleavage relative to the applied major principal stress, and Bayesian regression used to model peak strength as a function of confining pressure for each group. The analysis indicates that both variance of strength and the effect of confining pressure on strength are dependent on orientation. This suggests that strength characteristic values are not intrinsic properties, but are dependent on the loading state. This concept is currently not embodied within Eurocode 7, and so the paper concludes with suggestions for essential developments required to enable introducing strength anisotropy into limit state design.
Eurocode 7 (CEN 2004) (EC7) requires the characteristic value of a geotechnical parameter to be selected as a cautious estimate of the value affecting the occurrence of the limit state. Also, the Code requires selection of characteristic values for geotechnical parameters to be based on results and derived values from laboratory and field tests, complemented by well-established experience.
In the case of rocks with planes of anisotropy, such as foliation, cleavage or sedimentary structures, material strength has long been known to vary with the orientation of the applied major principal stress relative to the anisotropy planes (Jaeger 1960; Donath 1961). EC7 recognises this phenomenon, and notes that where applicable, it must be considered when assessing uniaxial compressive strength (UCS). EC7 is silent with regards triaxial strength.
In practice, it is only possible to test samples at a finite number of orientations, and the strength must be estimated for other orientations.
A review of the literature shows that numerous criteria and testing strategies have been proposed to capture the variation of strength with different loading orientations. On the basis of an extensive review, Duveau et al. (1998) classified the available criteria into the three distinct groups of mathematical continuous, empirical continuous, and discontinuous weakness plane models. More recent attempts to characterise strength anisotropy include those of Tien & Kuo (2001), Nasseri et al. (2003), Saroglou & Tsiambaos (2008) and Fjaer & Nes (2013). Although EC7 explicitly requires due account to be taken of anisotropy, no guidance is provided for either testing strategies or strength model selection.