Numerical simulations of the failure processes around openings in compressed rock specimens still compare poorly with observations. Neither a continuum approach in which the rock is assumed to have some effective plastic behaviour, nor a discrete approach in which fracturing is assumed to initiate and propagate according to certain fracture mechanical rules has led to satisfactory results. It is therefore concluded that both these approaches fail to represent the micro structure of rock in an appropriate way. An attempt is made to represent the micro structure of rock using a dislocation simulation program (DIGS) which is based on the displacement discontinuity boundary element method. Although progress so far has been limited, it is believed that this approach has the potential to reproduce and predict fracture processes in different rock types. This approach and the simulation of non-linear behaviour using plasticity theory are reviewed and compared in this paper.
Les simulations numeriques du developpement des fractures autour des ouvertures dans les specimens de roche comprime comparent pauvrement avec des observations. Ni une approche de continuite dans laquelle la roche a un effet plastique, ni une approche discrete, dans laquelle les fractures est supposee s'initier et se propager selon certaines reglements de mechanique des fractures, ont donne des resultats satisfaisant. II est done conclu que ces deux methodes ne representent pas la structure micro de la roche, dans une facon qui convient. Une tentative, utilisant un programme (DIGS) pour simuler des dislocations, qui est basee sur les deplacements des discontinuitees, dans la methode des elements de la bordure, età faites. Merne si le progres jusqu'ici a ete limite, nous croyons que cette approche oeuvre la possibility de reproduce et predire le developpement des fractures dans les roches differentes. Cette approche et la simulation du comportement non-lineaire, utilisant la theories de plastique, est retable et les resultats sont decries dans cue document.
Numerische Simulierungen der Bruchprozesse in der Nahe von Öffnungen von Gesteinteproben lessen such emmer notch schlecht mit tatsachlichen Beobachtungen vergleichen. Wider seine Kontinuumsannaherung, in der angenommen ward doss das Gerstein eon estimates effectives pastiches Verhalten besitzt, notch seine getrennte Annaherung, in der angenommen ward doss der Bruchprozess gemass bestimmter bruchmechanischer Regal stattfindet, haben zu befriedigenden Ergebnissen gefuehrt. Man schliesst darum aus diesen Beobachtungen, doss beide diese Stellungen misslingen, die Mikrostruktur des Gesteins entsprechend zu reprasentieren. In dieser Arbeit versucht man die Mikrostruktur des Gesteins innerhalb eines Verschicbungsdiskontinuitatsprogrammes (DIGS), doss von Grenzelementen gebrauch macht, zu reprasentieren. Obwohl Fortschritte diesbezuglich bisher begrenzt warren, glut man, doss diese Stellungsnahme das Potential hat Bruchprozesse in verschiedenen Felsarten zu reproduzieren und Vorherzusagen. Dieser annaherung und die Simulierung von nicht linearen Verhalten, doss von plastischer Theorie gebrauch macht werden in dieser Arbeit besprochen und vergliechen.
Fracturing and failure of rock is a complicated process of micro fracture growth and coalescence. In order to simulate the large scale effects of the fundamental micro processes within a numerical model it is often assumed that it is possible to capture the influence of the micro mechanisms by distributing them uniformly over a larger area. In a continuum approach such an area becomes a volumetric element, while in a large scale discrete fracturing approach that same area is populated by a set of discrete cracks. In order to calibrate the continuum models, cylindrical specimens of rock are generally subjected to uniaxial and triaxial loading configurations to obtain relations between applied stress and resulting strain, both before and after peak stresses have been reached. Although it has been pointed out by previous researchers (Read and Hegemier, 1984) that the observed brittle behaviour of rock does not represent the true material properties, but is a manifestation of the effects of progressively increasing inhomogeneity of deformation, this is not always fully appreciated. Read and Hegemier (1984) analysed uniaxially loaded specimens and concluded that the observed strain softening (brittle) behaviour was associated with a gradually reduced effective loaded area of the specimen due to progressive spalling of the specimen boundaries.
