Characterization of the thermo-hydro-mechano-chemical behaviour of uncemented reservoir quartz sand at high temperature and pressure

Resumen

Thermal Enhanced Oil Recovery (EOR) technologies, such as Steam Assisted Gravity Drainage (SAGD), Cyclic Steam Stimulation (CSS), Steam Flooding (SF) or Water Flooding (WF) use injection of steam or water in reservoirs to improve the oil recovery. Under certain conditions, such technologies activate complex Thermo-Hydro-Mechano-Chemical processes that affect the compressibility of the reservoir and the oil production, bring about environmental problems caused by subsidence, or jeopardise the integrity of the wells and surface facilities. The present research is part of a comprehensive feasibility study concerning EOR applications in an oilfield of heavy oil, located in the Orinoco Heavy Oil Belt (Venezuela). The reservoir consists of a solid skeleton formed mainly by uncemented quartz sands. In particular, this research has addressed, at macroscopic and particulate level, the compressibility of the soil skeleton (at effective stresses up to 50 MPa): Firstly considering the oil production alone; and secondly, when steam is injected (at temperatures up to 300ºC), as a consequence of the phenomena that arise. Heavy oil considerations are beyond the scope of this research. The rise of both temperature and intergranular stresses, caused respectively by the injection of steam and the release of pore pressure during oil production activate several phenomena that modify the porosity, permeability, and compressibility over time and have major effects on the macroscopic response (i.e. oil production) of the reservoir. At the particulate level, the following processes have been identified and studied: grain crushing, stress-corrosion, and pressure-solution. The effect of water composition, pH and temperature on compressibility, as compared with stress level alone, has been experimentally investigated. The work, funded by a partnership contract between the Repsol Technology Center (CTR) and the Technical University of Madrid (UPM), has been basically carried out at the CTR and the Department of Continuum Mechanics and Structures, Mechanical Computational Group, (UPM). A three-month stay at the Particulate Media Laboratory of the School of Civil and Environmental Engineering in the Georgia Institute of Technology (Atlanta), under the supervision of Professor J. Carlos Santamarina, has also been developed. The difficulty for testing representative undisturbed samples is undoubtedly a major drawback of laboratory characterisation, as sample recovery (mainly due to gas exsolution) and the handling cause inevitably disturbance of the original microfabric. However, leaving aside costly possible recovery techniques, purpose-built laboratory tests can provide some insight into the dominant factors affecting the sample response at the in situ conditions and, ultimately, cast some light on the prediction of the reservoir evolution. In this regard, the following series of uncoupled fundamental tests have proven to be clarifying in the understanding of the contribution of the processes involved (thermal, hydraulic, mechanical and chemical) in the phenomena at particle level mentioned above: Thermo-Hydro-Chemical tests: Six tests in an autoclave reactor for different residence times, pressures and temperatures (100, 200 and 300ºC) have been carried out. Oil-free sand quartz from the reservoir and a water solution with the same composition as the reservoir (brine) were employed for the experiments. As expected, the temperature and the composition of the water boost the dissolution of the quartz sand. In particular, Ca++ and Mg++ cations raise its dissolution since their higher coefficient of adsorption facilitates the interaction of the water molecules with the quartz surface. Thermo-Mechanical tests: Thirty-three oedometric tests on the oil-free sand at dry conditions have been carried out. The tests covered a broad range of effective stresses (0.5 to 50 MPa) and temperature (room temperature to 250ºC) and were monitored with acoustic emissions. The most unexpected conclusion drawn from the tests results is that temperature does not affect the compressibility of the samples. However, such a conclusion does not hold for the underlying microscopic processes. Apart from the rearrangement of particles, high stresses at low temperatures favour particle crushing, as inferred by the acoustic emissions and the comparison of the grain size distribution before and after the tests. As grain failure is negligible at high temperatures (250ºC), but grains subjected to high temperature show incipient fractures, the following processes have been assumed to yield strains: micro-fracturing of the textural features (asperities) of the quartz sands and subcritical cracking growth through stress-corrosion. The latter, assisted by the fluid inclusion in the detrital grains that act as pre-existing flaws and which water contribute to weakening its most overstressed region by the chemical rupture of the Si-O bonds. Thus, the fact that the compressibility remains unaffected with the temperature can be attributed to a change in grain-scale mechanics: from critical rapid crack growth, at low temperatures, to fluid-sensitive subcritical cracking, at the higher ones, without acoustic emissions. Once the previous series were completed and interpreted, a final series of fully coupled tests were undertaken: Thermo-Hydro-Mechano-Chemical tests: Six oedometric test and two core flooding tests have been carried out with oil-free sand at brine-saturated conditions, aimed at assessing the influence of the brine on the sand compressibility. The tests have been performed at a constant effective stress of 14 and 18 MPa, and 60 and 150ºC. It has been observed that the vertical strain is strongly dependent on the temperature and the renovation of the pore fluid, rather than on the effective stress. Hence, the vertical strain is re-activated right after the saturated-fluid is refreshed. Moreover, if a continuous flow is imposed, the vertical strain becomes greater and remains longer. A novel procedure aimed at keeping the fabric of samples after the tests for a detailed observation has been developed and patented, entitled “Fabric preservation of uncemented sand samples”. As a result, a detailed examination using computed tomography scan (CT-scan) and scanning electron microscopy (SEM) could be carried out. As expected, both stress-corrosion and pressure-solution features were identified. According to the tests, the activity of both of them is strongly dependent on the testing temperature. In particular, the stress-corrosion prevails in those tests carried out at lower temperatures, while stress-corrosion and pressure-solution coexist at higher ones, yet features of pressure-solution are more visible. These findings have been supported in the light of the comparison of the grain size distribution of the samples before and after the tests as well as of the determination of the silica concentration of the fluid collected at the outlet. Contrary to the expected, such concentration was negligible, and therefore it suggests that most of the silica dissolved was re-precipitated within the samples. Finally, a numerical model that accounts at a phenomenological level for pressure-solution has been implemented, as it has been found that it is ultimately the main process responsible for creep strains. The trend of the experimental and numerical data shows patterns with a reasonably good agreement so that, such matching underpins the fact that it is the prevailing process responsible for the observed creep behaviour at brine-saturated conditions and 150ºC.