Loading Events

« All Events

  • This event has passed.

Characterization of a CO2 leakage in the subsurface and its evolution from time-lapse seismic data

December 4, 2019 @ 2:30 pm - 3:30 pm

Stanislav Glubokovskikh

What to Expect

Injection of supercritical CO2 into aquifers reduces the elastic moduli of reservoir rocks, thus changing the seismic response. However, buoyancy-driven CO2 tends to form thin plumes with low-saturation edges and narrow streaks comparable in size to the seismic resolution (both vertical and lateral). Hence, accurate seismic characterisation and evolution forecasting of the plume are challenging. Otway Project Stage 2C (Victoria, Australia) was specifically designed to address these issues. In this talk, use the Otway experiment to examine capabilities and limitations of the surface seismic monitoring.

First, we set up a range of computer simulations, fluid flow and seismic, that explore key sources of uncertainty in the seismic image of the CO2 plume: data repeatability, rock physics of partial saturation and errors inherent in the inversion workflow. Secondly, the key findings from the synthetic and field inversion are used to establish a set of robust conformance metrics between history-matching fluid flow simulations and observed seismic anomalies. Even though the reservoir properties are known from multiple wells and the time-lapse response is relatively strong, the history matching based on reservoir simulations took more than a year and predict ±10% error in the CO2 mass and ±30% in the areal footprint.

Such a laborious workflow is impractical for short-term future prediction required for timely decision making in CO2 storage projects. Hence, we present an alternative framework for rapid forecasting of plume evolution, for use in conjunction with permanent seismic reservoir surveillance. Unlike a conventional physics-based simulator, the proposed algorithm predicts future seismic anomalies from the observed ones using artificial intelligence. The algorithm’s performance is illustrated using Otway data. The predictions are accurate for simple reservoir models and quickly adapt to more complicated scenarios.

Speaker Bio

Stanislav is a Senior Research Fellow at the Centre for Exploration Geophysics Curtin University (Perth, Australia). He got Diploma in geophysics from Dubna University (Russia) in 2008 and later received PhD in physical-mathematical sciences from Lomonosov Moscow State University (Russia) in 2012. In Russia, Stanislav worked for a number of commercial and government R&D laboratories until he joined the Curtin team at the beginning of 2015.

Stanislav’s research involves theoretical and practical aspects of seismic reservoir characterisation: all aspects of rock physics, quantitative interpretation of active/passive monitoring data. Stanislav serves as an associate editor of Geophysics.

Organizer

David Alumbaugh

Venue

LBNL, 74-104
1 Cyclotron Road
Berkeley, CA 94720 United States
+ Google Map