The effect of strain rate on inelastic strain development in porous sandstones deformed under reservoir conditions

Abstract

Fluid extraction from sandstone oil, gas, or geothermal reservoirs causes elastic and inelastic compaction of the reservoir, which may lead to surface subsidence and induced seismicity, as observed in the Groningen Gas Field, Netherlands. The inelastic compaction is partly caused by rate- or time-dependent processes, meaning that compaction may continue even if production is stopped. To reliably evaluate the impact of prolonged reservoir exploitation and post-abandonment behavior (>10–100 years), mechanism-based rate/time-dependent compaction laws are needed. We systematically investigated the effect of strain rate (rates of 10−3–10−9 s−1) in triaxial compression experiments performed on clay-bearing Bleurswiller sandstone (as an analogue of the Groningen reservoir sandstone) and almost clay-free Bentheimer sandstone, to explore the effect of mineralogy. Our results showed a systematic lowering of stress–strain curves with decreasing axial strain rate in Bleurswiller sandstone at differential stresses exceeding 40%–50% of peak stress (i.e. comparable to typical reservoir stress conditions). By contrast, in Bentheimer sandstone, rate effects were only noticeable at differential stresses > 70% of peak differential stress. Further investigation of the deformation behavior of Bleurswiller sandstone at varying confining pressure, temperature and pore fluid pH, complemented by microstructural analysis, suggested that the observed rate effects are likely controlled by rate-dependent intergranular frictional sliding at lower differential stress, with an increased role of stress corrosion cracking at higher stress. Extrapolation of our data to reservoir conditions suggests that additional strains of about 10% can be expected, compared to the strain accumulated at laboratory strain rates. Our results show that time-dependent inelastic deformation plays an important role in controlling reservoir deformation, such as of the Groningen gas reservoir. Such effects could lead to an underestimation of surface subsidence and induced seismicity, if not accounted for. The present experiments provide important data for developing physics-based constitutive models for predicting rate/time-dependent reservoir compaction.