This article identifies and discusses the scientific challenges of hydrogen storage in porous media for safe and efficient large-scale energy storage to enable a global hydrogen economy.

No abstract

Pore pressure reduction in sandstone reservoirs generally leads to small elastic plus inelastic strains. These small strains (0.1%-1.0% in total) may lead to surface subsidence and induced seismicity. In current geomechanical models, the inelastic component is usually neglected, though its contribution to stress-strain behavior is poorly constrained. To help bridge this gap, we performed deviatoric and hydrostatic stress cycling experiments on Slochteren sandstone samples from the seismogenic Groningen gas field in the Netherlands. We explored in situ conditions of temperature (T = 100°C) and pore fluid chemistry, porosities of 13% to 26% and effective confining pressures (≤320 MPa) and differential stresses (≤135 MPa) covering and exceeding those relevant to producing fields. We show that at all stages of deformation, including those relevant to producing reservoirs, 30%-50% of the total strain measured is inelastic. Microstructural observations suggest that inelastic deformation is largely accommodated by intergranular displacements at small strains of 0.5%-1.0%, with intragranular cracking becoming increasingly important toward higher strains. The small inelastic strains relevant for reservoir compaction can be described by an isotropic, Cam-clay plasticity model. Applying this model to the depleting Groningen gas field, we show that the in situ horizontal stress evolution is better represented by taking into account combined elastic and inelastic deformation than it is by representing the total deformation behavior using poroelasticity (up to 40% difference). Therefore, inclusion of the inelastic contribution to reservoir compaction has a key role to play in future geomechanical modelling of induced subsidence and seismicity.

a b s t r a c tGeological storage of CO 2 in clastic reservoirs is expected to have a variety of coupled chemical-mechanical effects, which may damage the overlying caprock and/or the near-wellbore area. We performed conventional triaxial creep experiments, combined with fluid flow-through experiments (brine and CO 2 -rich brine) on samples of poorly consolidated, carbonate-and quartz-cemented Captain Sandstone from the Goldeneye field. The main goal was to study the effect of carbonate cement dissolution on mechanical and ultrasonic properties, as well as on the failure strength of the material. Our experiments were performed under in situ reservoir conditions, mimicking reservoir depletion and injection. Although total dissolution of calcite was observed, and confirmed by microstructural and fluid chemistry analyses, it did not affect the rock mechanical properties, nor was any measurable rock strength reduction observed. This is most likely because grain-to-grain contacts were sufficiently quartz-cemented and quartz is not affected by CO 2 -rich brine. Failure data for the Captain Sandstone showed that the stress conditions under which CO 2 injection will take place remain far away from the failure envelope. Therefore, CO 2 injection is not expected to lead to shear failure of the reservoir. However, longer-term chemical reactions, involving minerals such as feldspar, clays or micas, still require more research.

Storage of anthropogenic CO2 in geological formations relies on a caprock as the primary seal preventing buoyant super-critical CO2 escaping. Although natural CO2 reservoirs demonstrate that CO2 may be stored safely for millions of years, uncertainty remains in predicting how caprocks will react with CO2-bearing brines. This uncertainty poses a significant challenge to the risk assessment of geological carbon storage. Here we describe mineral reaction fronts in a CO2 reservoir-caprock system exposed to CO2 over a timescale comparable with that needed for geological carbon storage. The propagation of the reaction front is retarded by redox-sensitive mineral dissolution reactions and carbonate precipitation, which reduces its penetration into the caprock to ∼7 cm in ∼105 years. This distance is an order-of-magnitude smaller than previous predictions. The results attest to the significance of transport-limited reactions to the long-term integrity of sealing behaviour in caprocks exposed to CO2.

No abstract

[1] Geological storage of CO 2 in clastic reservoirs and aquifers is expected to have a variety of coupled chemical-mechanical effects. To investigate the effects of CO 2 injection on creep phenomena, we performed uniaxial compaction experiments on granular aggregates of quartz and feldspar under both wet and dry control conditions. The experiments were performed in constant stress mode. Grain size, temperature, CO 2 partial pressure, and effective stress were varied in order to determine their individual effect. Pore fluid pH was varied by the injection of CO 2 and by addition of acidic and alkaline additives. Pore fluid salinity was increased by the addition of NaCl. Wet samples showed instantaneous compaction upon load application, followed by time-dependent creep. From the mechanical data and microstructures, the main compaction mechanism was inferred to be chemically enhanced microcracking in both quartz and feldspar, with subcritical crack growth, i.e., stress corrosion cracking, controlling deformation in the creep stage. The injection of CO 2 and the concomitant acidification of the pore fluid inhibited microcracking in both the quartz and feldspar samples in line with known effects of pH on stress corrosion cracking. We infer that the injection of CO 2 into quartz-and plagioclase-bearing sandstones will inhibit grain scale microcracking process and that related geomechanical effects, such as reservoir compaction and surface subsidence, will be negligible compared with the poroelastic response.

Reduction of pore fluid pressure in sandstone oil, gas, or geothermal reservoirs causes elastic and possibly inelastic compaction of the reservoir, which may lead to surface subsidence and induced seismicity. While elastic compaction is well described using poroelasticity, inelastic and especially time‐dependent compactions are poorly constrained, and the underlying microphysical mechanisms are insufficiently understood. To help bridge this gap, we performed conventional triaxial compression experiments on samples recovered from the Slochteren sandstone reservoir in the seismogenic Groningen gas field in the Netherlands. Successive stages of active loading and stress relaxation were employed to study the partitioning between elastic versus time‐independent and time‐dependent inelastic deformations upon simulated pore pressure depletion. The results showed that inelastic strain developed from the onset of compression in all samples tested, revealing a nonlinear strain hardening trend to total axial strains of 0.4 to 1.3%, of which 0.1 to 0.8% were inelastic. Inelastic strains increased with increasing initial porosity (12–25%) and decreasing strain rate (10−5 s−1 to 10−9 s−1). Our results imply a porosity and rate‐dependent yield envelope that expands with increasing inelastic strain from the onset of compression. Microstructural evidence indicates that inelastic compaction was controlled by a combination of intergranular cracking, intergranular slip, and intragranular/transgranular cracking with intragranular/transgranular cracking increasing in importance with increasing porosity. The results imply that during pore pressure reduction in the Groningen field, the assumption of a poroelastic reservoir response leads to underestimation of the change in the effective horizontal stress and overestimation of the energy available for seismicity.