[en] This book is a special issue of the journal 'Revue francaise de genie civil' (vol.6, no 1/2002). It gathers several papers written in the framework of the CNRS-ANDRA 'Forpro' research program on the feasibility of radioactive waste disposal in deep geological formations. It treats of the researches carried out on the fracturing of granites and argilites: physical and mechanical characteristics of the Vienne granite and of the Argilite de l'Est formation; modeling of the anisotropic damaging by meso-cracking of the Vienne granite; modeling of the anisotropic damaging of the Vienne granite; microstructural modeling of the Vienne granite damaging; modeling of an argilite using the CLoE model; elasto-plastic modeling coupled with the induced anisotropic damaging for the argilites; comparison between model concepts and forecasts. (J.S.)

No abstract available

No abstract available

No abstract available

No abstract available

[en] The evolution of crack growth velocity v vs. stress intensity factor K in Meuse/Haute-Marne argillite is experimentally investigated at different relative humidities. A new algorithm based on finite elements simulations aimed to process the measured data is proposed and allows the calculation of the v(K) curves. It is observed that the crack velocity and the stress intensity factor respectively increases and decreases as the relative humidity increases. (authors)

No abstract available

[en] Document available in extended abstract form only. The double torsion (DT) test is used here to characterize the evolution of subcritical crack growth. It consists in the bending of a specimen having an initial notch and which lies on two supports parallel to the crack. More precisely, the load is applied on both sides by a hydraulic press (capacity 20 daN, precision 0.01 N) while the displacement of the pump is measured by two LVDT sensor (precision 0.01 mm). To detect the initiation of the crack and to measure its length during the relaxation phase, a numerical camera (resolution 25 μm) has been positioned under the specimen. The typical dimensions are: L = 120 mm, d = 3 mm/dn = 2 mm, W = 60 mm, Wn = 17.50 mm, a = 20 mm. The argillite specimens are partially saturated to investigate the influence of hygrometry on the tenacity and crack velocity. Once the hydric equilibrium is reached, the test can be carried out, and consists in two steps: - a rapid loading phase to reach the initiation of the crack, - a relaxation phase to analyze the crack growth. The relative humidity (RH) is kept constant thanks to a device controlling the circulation of air. For each tests, the stress intensity factor (mode 1) K_1C is computed from: K_1C P_c.W_n.[3(1+?)/(W.d³.d_n)]^1/2, with P_c the maximal load and? the Poisson's ratio. Furthermore, the initial crack length (L_i) and final crack length (L_f) are determined by analyzing the recording of the numerical camera. Following a similar approach as in Ciccotti et al., 2000-2001, the crack velocity ν = L is determined by derivating the relation k = P/y (with k the stiffness of the specimen and y the displacement) with respect to time. If the function k(L) is calculated by numerical simulations, and as y is kept constant during the relaxation, one obtains: V = L x P / (y dk/dL) These experimental results show the influence of hygrometry on subcritical crack growth rate: the final crack length seems to decrease with the increase of hygrometry. Besides, for the upper hygrometries in 85%-90%, this reduction seems very important. (authors)

[en] A modified elasto-damaged model counting for laboratory observed mechanisms is presented. Even if the general framework of the model is clearly phenomenological one, laws of state variable evolution are inspired from micro-mechanical mechanisms of crack propagation and they are constructed upon fracture mechanics bases. Basic hypothesis and the determination procedure of model parameters. The model is compared to the test laboratory results. (author)

[en] The microstructural approach in damage modeling, which is presented in this paper describes the evolution of micro-crack geometry as a function of history loading. If the crack geometry is known, the effective properties could then be calculated foe any cracked rock by the mean of a micro-mechanical model. The P_L evolution law which is necessary in the describing of crack geometry evolution is hardly based on the crack microscope observation as well as on the theory of fabric tensors. This approach is applied in the modeling of mechanical behaviour of Vienne granite. The result of model simulations are compared with laboratory tests. (author)