Stress Path and Hysteresis effects on Integrity of CO2 Storage Site

The aim of the SPHINCSS project was to further investigate the risk picture associated with large-scale CO2 storage, especially about plans offshore Norway. The focus has been on geomechanical impacts on the ability of a selected storage reservoir to keep CO2 safe in the rock formation. Geomechanics is the study of pressure distribution in a rock that is influenced by the interaction between the pore fluid pressure and solid pressure in three directions that is transferred between the rock grains and mineral cement between neighbouring grains. Any increase or decrease in the pore pressure due to injection or production of fluid (in our study CO2, brine or oil and gas) will affect the solid pressure components, called stresses. Both reservoir and caprock stresses are affected, especially at boundaries between the two and where faults occur between them. In the project we studied in detail cores from Draupne shale (donated by Equinor) and evaluated its ability to withstand exposure to CO2. This was done by measuring strength and stiffness changes when a plug is exposed to many variations in CO2 and saline mixtures. In this project, we introduced new methods to investigate the impact of CO2 on mechanical shale properties: - Long compression tests on small cores with the SINTEF miniature pressure cell, where a small hole is drilled in the centre of the plug for faster penetration of CO2 dissolved in brine. In these tests stiffness and strength are measured without CO2 and the same procedure is tested with exposure to CO2. Here we saw some weakening of the plugs exposed to CO2, but to a small extent, probably because Draupne does not contain much acid-sensitive minerals. - Low frequency testing, where shale plugs are subjected to cyclic compression in a special cell, designed to replicate the effect of a seismic wave hitting the plug. Here the deformation is so small that it does not destroy the plug, and therefore the test is such one can test on the same plug the effect on the rigidity of reference saline saturation with the addition of CO2 later in the test. Stiffness is measured continuously with very high precision, and therefore longer trends can be identified. Here we saw negligible change in the stiffness of the plug when CO2 was introduced. - "Punch" testing, where a specially-designed compression cell is used on 5 mm diameter "pancakes" of Draupne shale. A piston punctures a small hole in the disc with the needed force being measured. The shear strength of the material can then be calculated. Many such tests can be performed on the same plug, due to the small dimensions needed. The Draupne shale was exposed to brine, CO2 dissolved in brine, supercritical CO2 dissolved in brine, dry CO2 and air. The tests were repeated for each exposure after 1 day, 2 days and 7 days to show time evolution of the shale strength. Here we saw that dry CO2 (in the field, dry CO2 is injected, but it picks up moisture the longer it is in contact with pore water) weakens less the shale than CO2 dissolved in brine, which weakens the stone a little more than only brine. Other laboratory activities included examining natural fractures in Draupne shale using shear box testing at NGI in Oslo and using UiO's HADES cell under synchrotron radiation in Grenoble. The shear box displaces one side of a fracture relative to the other which remains stationary, and friction is measured as a function of velocity. The results of this testing are that compression occurs during cutting, which indicates that a crack or fault in Draupne would heal during deformation, i.e. good news. Initial analyses from the HADES test show that one can identify CO2 drops that penetrate into a crack (which we induced with compression testing), and that longer exposure can show whether the crack opens more or heals and closes, since we managed to calculate all local deformations using image analysis. On the modelling side, we were able to use SINTEF's MDEM fracturing code to simulate the risk of leakage along a fault. Results show that reduction in pore pressure due to oil production can create cracks along the fault and thus a leakage path if CO2 is injected into the same reservoir, while an aquifer can withstand more pressure increase. These conclusions were verified with a commercial code. Another simple model simulated fracture risk in the reservoir as a result of deposition of dissolved particles in the pore throats and increased injection pressure. As a Researcher Project, it was also important to focus on education: a post-doc and 3 French Master students participated in laboratory and modelling work, which resulted in 2 journal articles. In addition, Post-doc N. Agofack travelled twice to France to visit A. Bois from Curistec in Lyon and Prof. J. Sulem from Ecole des Ponts in Paris who were advisers in the project and assisted in planning and interpreting the experiments. N. Agofack also spent 3 weeks at the Lawrence Livermore lab in California.

Prosjektet har undersøkt potensielle forskjeller mellom et tømt olje-reservoar og en akvifer med uendret opprinnelig poretrykk. En akvifer bygger raskere opp trykk, mens en begrenset injeksjon av CO2 i et depletert reservoar vil ikke overstige opprinnelig trykk. Likevel viser våre simuleringer at å tømme olje eller gass fra reservoaret kan forårsake irreversible spenningskonsentrasjoner på forkastninger langs reservoaret. Når poretrykket øker igjen, vil det kunne lekke CO2 langs forkastningene, mens skadene for akviferer er betydelig lavere. En virkning vil derfor være at mer modellering og karakterisering bør anbefales når et større lagrinssted vurderes og at forkastninger må studeres i større detalj enn i dag. Samtidig, peker prosjektet på at skiferformasjoner ikke er like og sannsynligheten for å forårsake skjærbrudd i toppbergart er svært liten. Resultater fra dette prosjektet er med å øke tilliten til CCS generelt og bidrar til vurdering av CCS i større skala enn pilotskala.

In order to ensure CO2 storage site containment, the highest confidence in modelling the geomechanical changes occurring in the surrounding rocks has to be attained. Prior work has highlighted that expected stress changes will be of different nature at different locations in a caprock overlying a storage reservoir and in the proximity of faults bounding it. In addition, different stress paths are expected for aquifers and depleted reservoirs, again impacting expected stress distribution in the caprock. Hysteresis here refers to non-uniqueness in explored deformation states upon varying stress; these stress variations can be the outcome of pressure cycles in the reservoir, geological stress variations (burial and uplift sequences) and in the near-well area, the consequence of cyclic injection and temperature values. This project will explore the consequences of varying pore fluid pressure in a CO2 storage site on the stress state in the reservoir (particularly around the injection well), the caprock and bounding faults. This will be done for aquifers and depleted oil/gas fields, for features taken from North Sea candidate sites proposed by Statoil. - Experiments will mimic fluid injection and well shut in cycles, monitoring fatigue effects and thermal stress possible damage. Exposure tests to CO2 will complement the investigation, addressing geochemical effects on rock stiffness and strength, for several lithologies. These tests will be conducted on state-of-the-art rigs at SINTEF and Ecole des Ponts Paris and analysed in collaboration with Curistec, France and Lawrence Livermore National Laboratories, USA. - Stress path effects on rock specimens including field shale donated by Statoil will be measured in triaxial conditions and for partial CO2 saturation, combined with seismic and ultrasonic acoustic measurements at SINTEF. - Fracture shear and fault reactivation tests will be carried out at NGI. NGI and SPR will then run reservoir scale geomechanical modelling.