ABSTRACT
Observations and studies of the mechanical behaviour of solid rocks and rock masses play a fundamental role in determining the stress conditions during and after the creation of e.g. underground structures, bridges, geothermal drilling or cable lines. For example, responses of buildings and structures can be observed with photogrammetric measurement techniques (Albert & Seyler 2004). In contrast, for determination and monitoring of structure-rock and rock-rock interactions, installations of established instruments for the characterization of the mechanical behaviour (e.g. extensometer, inclinometer) are usually required. Alternatively, the mechanical behaviour of rock masses and hard rock samples can be detected with the finite deformation analysis. This is a method of visual quantification of the micro-scales deformation behaviour of rocks (Dunnet 1969, Dunnet & Siddans 1971, Fry 1979, Mulchrone 2007) under compressive and extensional stress regimes and can be carried out without installation of measuring instruments on for example inner walls of underground cavities. The finite deformation analysis utilises the fact that
rigid particles tend to adjust themselves in viscous materials. Under mechanical stress this results in orientation and deformation of the particles in the rocks (Ramsay & Huber 1983, Mulchrone 2007). Ramsay (1967) came up first with the idea of measuring finite strain from randomly oriented particles, once with the Rf/φ-Method and second with the object-object separation method, which is known as the Fry-Method
(Fry 1979). The different types of the Fry-Method, developed byErslev (1988) andErslev&Ge (1990), as well as the Rf/φ-Method are only applicable for elliptically shaped particles. A Normalized Fry-Method was modified by McNaught (1994) to incorporate the grain shape in the deformation analysis. Förster & Lempp (2013) show certain problems of the softwareimplemented analysis of the Normalized Fry-Method after McNaught (1994). A combination of the Rf/φMethod and of the Fry-Method types avoids the problem of the competence contrast between the particles and the matrix. In addition to the combination of these two methods, there is also the Intercept-Method (Saltikov 1958, Hilliard 1962, Launeau et al. 1990, Launeau & Robin 1996, Launeau & Robin 2003), in which the mechanical difference between particles and matrix is not necessary. The Intercept-Method is based on the determi-
nation of the particles shape preferred orientation (SPO) by measuring the shape of deformed particles. Also the mean elongation of irregularly shaped particles is transferred into a non-random orientation (Blumenfeld & Bouchez 1988, Allard & Berm 1989, Shelley 1993). Launeau & Robin (2003) developed the software tool SPO2003 in which the InterceptMethod is implemented for a digital and objective evaluation of anisotropic distributed and irregularly shaped particles.
