The Use of Distributed Acoustic Sensing for 4D Monitoring Using Vertical Seismic Profiles: Results from the Aquistore CO2 Storage Project

Abstract

2017 ©Copyright Kyle Harris, 2017 ii 1. Abstract Distributed acoustic sensing (DAS) is a rapidly developing technology that employs optical fibers to detect acoustic disturbances in the subsurface. The Aquistore CO2 storage project uses time-lapse DAS VSP's to monitor injected CO2 in a deep geological reservoir (>3200 m) beneath the study site near Estevan, Saskatchewan, Canada. Reservoir monitoring is a crucial component to any geological storage project, and aims to ensure the safe and effective containment of the injected fluids, and evaluate the integrity of the storage and sealing units. An assessment of the ability of DAS VSP's to detect the CO2 response has been conducted for the Aquistore site. Prior to injection, fluid flow simulations were performed to predict CO2 distributions in the reservoir. These simulations were used to model the response of the CO2 plume in 2D time-lapse DAS VSP's. The expected responses were compared with estimates of time-lapse noise computed with field data from the baseline survey, and it was demonstrated that the plume would be visible in the reservoir after 27 kt of injection. After 36 kt of CO2 injection, the first monitor dataset was acquired. 4D VSP processing and imaging were applied to produce time-lapse difference volumes of the reservoir. Acceptable repeatability was attained, with normalized root-mean-square (nRMS) values less than 0.4 beneath the observation and injection wells. An anomaly in the lower part of the reservoir near the observation and injection wells was attributed to the replacement of brine with CO2.iii Updated fluid flow simulations were obtained that better replicated the injection parameters observed at the study site. Forward seismic modeling was then performed for 36 kt and 97 kt injection scenarios that reflected the 3D shot and receiver geometry in the field acquisitions. These data were processed using the same sequence applied to the field data to obtain comparable time-lapse difference volumes. Comparisons between the field and synthetic 36 kt anomalies were used to refine interpretations of the CO2 distribution in the reservoir. However, they also revealed the need to update the geological model to better reproduce the laterally asymmetric plume expansion.iv