Abstract
This report summarises the work done on understanding the physical mechanisms and mechanics governing compaction behaviour of the Slochteren sandstone, performed within a NAM-funded research programme (2015-2021) focussing on the behaviour during production, and within the DeepNL research programme (2018-present) focussing on the post-abandonment phase. In these collective studies, microstructural
... read more
observations, obtained from field material (depleted and undepleted core from the Groningen Gas Field) and experimental work (conventional triaxial stress and uniaxial strain experiments), combined with experimental mechanical data, simulating stress changes during production of the Groningen Gas Field, enabled us to identify the main grain-scale deformation mechanisms operating in the Slochteren sandstone. The Slochteren sandstone mainly consists of quartz grains, and smaller amounts of K-feldspar, with the grain-to-grain contacts frequently containing μm-thick, intergranular clay films. Experimental and microstructural work suggested that most of the compaction observed in the Slochteren sandstone developed near-instantaneously, i.e. over time-scales of hours-days. Slower experiments performed over weeks to months showed decelerating creep deformation, contributing a modest 10-20% to the inelastic strain accumulated during active loading, i.e. a 10-20% time-dependent contribution to strain. Together with the mechanistic constraints of the field, this suggests that reservoir compaction in the Groningen Gas Field is largely rate-insensitive, with only a modest contribution of time- or rate-sensitive mechanisms. During production, compaction is dominated by (virtually) rate-insensitive processes, like poro-elastic deformation, and compaction of and slip along the intergranular clay films. On longer time-scales of weeks-months, time-dependent compaction may be governed by a different mechanism, notably by stress corrosion (intragranular) cracking. To what extent even longer-term compaction behaviour (decades-centuries) will be influenced by this slow creep process, or by other creep processes such as pressure solution, still requires further investigation and quantification. Extrapolating experimental data obtained at the slowest strain rates achievable in the lab (i.e. 10-8 or -9 s-1) to the current compaction rate in the field (i.e. 1000 slower) was attempted using various empirical rate-dependent models, such as the RTCM model. Doing so suggested that additional strains of 10-50% can be expected (i.e. inelastic strain in excess of what has already been accumulated at rapid/lab deformation rates). Though experimental work hints at excess strain in the field after abandonment likely being at the lower end of this range (10-20%), the broad range in forecasted strain illustrates the need to better constrain the models used for extrapolation to field timescales. Since such slow deformation rates cannot be accessed in the lab, understanding the microphysical processes underlying sandstone deformation is required. Microphysical models describing rate-insensitive compaction were implemented in Discrete Element models to assess sandstone compaction behaviour at the cm-dm scale. These numerical models can be used to evaluate reservoir compaction in different locations on the field due to pressure equilibration or repressurisation, with rate-sensitive mechanisms, such as stress corrosion cracking, to be added at a later stage. However, it should be noted that even if the Slochteren reservoir formation would not exhibit much ongoing compaction after field closure due to time-dependent compaction processes, that does not necessarily imply independence of subsidence and seismicity from production rate and strategy. Even if reservoir compaction is fully time-independent, subsidence and seismicity may be influenced by time-dependent (creep) deformation of the overlying Basal Zechstein and rock salt, and by pore pressure re-equilibration within/near the reservoir (e.g. the clay-rich Ten Boer and Ameland members), and/or the underlying low-permeable (Carboniferous) shales. In addition, time dependent behaviour may perhaps be caused by transient fluid flow along faults.
show less