Cross-well Seismic Tomography to Monitor CO2 Flood Area Using Difference Analysis Method

Capturing CO2 directly from large stationary sources such as thermal power plants and subsequently storing it in a nearby aquifer can be the most efficient way to reduce CO2 emissions into the atmosphere at a lower cost, and monitoring injected CO2 is important to verify practical effectiveness of the CO2 storage. The pilot-scale CO2 sequestration experiment has been undertaken at the Nagaoka gas and oil field, in Niigata Prefecture, Japan and we conducted time-lapse cross-well seismic tomography to detect the spread of injected CO2. When super- critical CO2 spreads into porous media saturated with brine water, the seismic velocity always decreases, which is confirmed theoretically and experimentally. Therefore, CO2 flood area can be estimated from seismic tomography records before and after CO2 injection. However, the tomography records before and after CO2 injection include not only the variation due to CO2 flood, but also differences occurred from the location and property of sources and receivers and they result analysis errors. Thus, in this study, we applyed the differential analysis to cross-well seismic tomography for reducing the difference due to measurement conditions and getting more accurate monitoring results. In the normal procedure, velocity differences are calculated after applying inversion processing to each record separately, but in this differential analysis, first forward modelling is applied using the velocity distribution inverted from one referential record, second the set of first arrival time is obtained from the modelling result and finally the inversion processing is analyzed from the data set of the first arrival time added with the difference of travel-times from the referential record. In the result, we can have an analysis result in which the distribution of CO2 flooding is clearly identified along with cap rock under the inversion analysis with the constrained condition of no velocity enhancement.