| Summary: | The injection of Carbon Dioxide (CO2) into saline aquifers could help to reduce anthropogenic CO2 emissions, but due to the potential dangers posed by leakage, extensive monitoring will be required. Ideally this monitoring will be able to determine: CO2 movement, the amount of CO2 stored, the CO2 phase (gaseous, supercritical or dissolved in brine), the trapping mechanism (structural or residual) and the occurrence of pore-pressure changes. The literature review highlighted five uncertainties which could impact on the ability of time-lapse seismic surveys to fulfil the monitoring requirements: the dependence on reservoir properties, the fizz gas effect, the CO2 phase, uncertainty regarding the appropriate fluid distribution model and the effect of pressure build-up. An integrated geological, rock physics, seismic and fluid-flow modelling approach was developed to predict the seismic response to CO2 injection. Two approaches were taken, the first excluded fluid-flow simulation and predicted the sensitivity of the seismic response to reservoir changes, the second included outputs from fluid-flow simulation to predict the time-lapse response. The majority of the modelling was performed for a synthetic clastic saline aquifer at 1500m depth. The sensitivity study was applied to all the uncertainties, whereas fluid-flow simulation was only included to examine fluid distribution and pressure build-up. Through examination of these issues, the ability of the time-lapse seismic surveys to achieve the monitoring aims was established. The zero-offset reflectivity was sensitive enough to detect the migration of both structurally and residually-trapped CO2 irrespective of the fluid-distribution and reservoir depth (up to 2500 m). Unfortunately there was no success with establishing a seismic monitoring approach which could determine the CO2 saturation when a homogeneous fluid distribution was assumed. With regards to detecting the CO2 phase and trapping mechanism there were mixed results. The seismic response was not sensitive enough to distinguish brine from brine saturated with CO2, however the gaseous and supercritical CO2 could be differentiated by AVO cross plotting. The potential for a reflection of the interface between residually and structurally trapped CO2 was identified; this reflection could help constrain how much CO2 was retained via these different mechanisms. Overpressure could be identified by anomalously large amplitude changes and pushdown both within the plume and away from the plume on the zero-offset reflectivity. Pressure and saturation changes could be distinguished from each other using the AVO response. This research suggested a link between the trapping mechanism and fluid distribution model, which could mean that the appropriate fluid distribution model for seismic interpretation may vary over the life-time of a sequestration project. These results were used to guide the development of a monitoring framework for sequestration sites, which could be tailored base upon the specific requirements of the site. Overall conducting this modelling study highlighted the advantages of incorporating fluid-flow simulation results into a seismic sensitivity study and underlined the importance of conducting a monitoring sensitivity study prior to site selection.
|
|---|
