Crack propagation in reservoir rocks in the CO2 storage context
Suhett Helmer, Gisele
MetadataShow full item record
The CO2 capture and storage (CSC) in the deep geological formations is a promising solution to decrease the undesirable effects of the CO2 concentration in the atmosphere. One of the most important aspects concerning the public acceptability of this technology and its use in the industry is the demonstration of the storage safety and integrity. The geological formations, being very heterogeneous environments, often have crack networks. The environment hydromechanic response under the combined effect of the pore pressure due to the injection and the chemical reactions that generate precipitation and dissolution of different minerals is a key parameter to study the long term storage integrity. The combined effect of the effective stress variation and the chemical reactions can provoke the opening, the clogging and/or the crack propagation of pre-existent cracks. The fracture toughness (KC) is a parameter associated with the ability of the material to resist the crack propagation. The crack propagation can be due the change of the state of stress (mechanical effect of the injection), or the rock fracture toughness change due to the rock degradation (chemical effect). Knowing the rock fracture toughness and its evolution due to the chemical effects it’s then, important to the modeling of crack propagation in the geological CO2 storage context. The two principal objectives of this work are: (1) The experimental evaluation of the CO2 degradation over the reservoir rock fracture toughness (to understand the mechanism in a smaller scale) (2) The reservoir integrity evaluation due to the CO2 injection (to understand the mechanism in a bigger scale) Concerning the experimental program performance, one preliminary phase consists on the choice of the study rock, the degradation mode and the most appropriate laboratory tests to reach the study objectives. Thereby, a limestone (Pierre de Lens) is chosen to be study both in a sound and after being degraded by the CO2. The rock degradation is made in a autoclave and the samples are placed in a aqueous solution saturated with CO2, under reservoir conditions (temperature of 60 °C and pressure of 15 MPa) during a degradation time of a month. Many configurations for the mechanical laboratory test were chosen for testing the fracture toughness of the rock both on mode I and mode II of crack propagation. To the crack propagation on mode I (opening): - The Brazilian test (BDT) ; - The Central Crack Notched Brazilian test (CCNBD -- half-moon crack format) and the Central Crack Brazilian test (CCBD – rectangular crack); - The semi-circular three point bending test (SCB); - Pour la propagation d’une fissure en mode II (cisaillement dans le plan de la fissure) : - The semi-circular three point bending test with asymmetric supports (ASCB); The Central Crack Notched Brazilian test (CCNBD -- half-moon crack format) and the Central Crack Brazilian test (CCBD – rectangular crack); - The Punch Through Shear Test (PTST) performed in a triaxial cell. Some of the mechanical tests will be performed using a digital image correlation technique (DIC). This device allows evaluating the rock fracture toughness from the displacement fields on the samples in parallel to the classic methods based of the rupture load (of peak load). The utilization of the image correlation technique it was made under the monitoring of the LMT-Cachan (Mechanic and Technology laboratory of the Normale Supérieure de Cachan). It is important to precise that a part of the tests (most of them) were realized on ambient humidity. Another part aim to study crack propagation under the effect of the water presence and the presence of and water with CO2 dissolved under a mechanical load. As a complement of the mechanical tests, a microstructure characterization of the rock was made and its evolution before and after degradation was realized using observations in a scanner electron microscope (SEM). It was also made a mercury porosimetry analysis. The experimental results show that the fracture toughness values for the mode I obtained by the different types of tests and analyzed by the different methods are concordant. The image correlation technique put in evidence the fact that the evaluation for the mode II crack propagation mode can’t be evaluated on non confined tests. Two degradation procedures were carried on : (1) the samples were put in a CO2 saturated water in an autoclave ; (2) the autoclave water is renewed each week. For each case we could note that the porosity didn’t change much (0.4%). This variation is equivalent to the one obtained by numerical simulations for a limestone reservoir for a 10 years period in a zone far from the injection point (zone of CO2 saturated water). The SEM analyses show an observable degradation due to a variation on the porosity distribution before and after the degradation. The rock fracture toughness it’s not much affected by this variation passing from 0.62 to 0.58 MPa.m0,5. The effect of water presence being also studied, mechanical tests were performed on saturated samples (submersed). This has a great influence over the fracture toughness parameter with a decrease of 17% (comparing with dry samples).This results it not much affect by the presence of CO2 on the fluid. Concerning the analysis on mode II, many tests were made (in a triaxial cell), with samples under different confined pressures (5, 10 and 15 MPa). It was possible to notice that the PTST test allows a great evaluation of the fracture toughness on mode II (3 MPa.m0.5 for the Pierre the Lens). Nevertheless, the mode I is still present for the confinement pressure of 5 and 10 MPa and it’s not always inexistent for the 15 MPa pressure. It was also shown that the CO2 influence is not important, with a fracture toughness passing from 2.96 to 2.77 MPa.m0,5.Another important point is the numerical modeling concerning the influence of the sample crack thickness that favor the appearance of the mixed mode. Its influence decreases with the increase of the confinement pressure. The influence of the rock degradation by the crack propagation was also studied by a numerical modeling using the numerical model ENDO-HETEROGENE, developed by the BRGM (in the context of the Nicolas Guy PhD – coordinated both by BRGM and LMT-Cachan) and integrated on the finite elements code – Code-Aster®. Its model is based on the crack initiation and crack propagation in a heterogeneous environment where the material parameters variability follows a two parameters Weibull distribution. It was explored the possibility that the chemical degradation that reduces the fracture toughness will also influence the Weibull parameters changing the heterogeneity of the crack distribution on a microstructure scale (pre-existent defects). Its heterogeneity is represented on the Weibull distribution by the parameter m. Thereby the impact of these variations (on the fracture toughness Kc and the parameter m) on the range of the data obtained by the experimental study it was study with the aid of the numerical model to analyze the crack initiation and the crack propagation on a sedimentary layer on reservoir conditions. The results show that the parameter m can influence the number and the dimensions of the cracks but it does not change the maximum size of the crack. Putting the experimental results in context with the studied rock, the Weibull parameters were evaluated for the sound and the degraded rock. It was observed that m changes from 8.55 to 8.52 and σ0 from 2.8 to 2.2 MPa (sound and degraded rock respectively). The numerical results show that these variations are not enough to change the crack network created after a load under a geological layer. This results show that in the case of a limestone reservoir the CO2 injection won’t influence significantly neither the fracture toughness nor the probabilistic parameters. These results correspond to a period of 10 years of injection in a zone far from the injection point.