Application of Full Waveform Inversion to Time-Lapse Walkaway Vertical Seismic Profile Data for CO2-EOR Monitoring in Abu Dhabi: A Feasibility Study on Field and Synthetic Data

This study aims to investigate the feasibility of CO2-EOR monitoring by full waveform inversion (FWI) of time-lapse VSP data in an onshore CO2-EOR site in Abu Dhabi. CO2-EOR monitoring using conventional time-lapse surface seismic in onshore oil fields in Abu Dhabi is often technically challenging for two main reasons. The first is that elastic property change in response to pore fluid substitution is relatively small because the elastic modulus of the reservoir rock frame is far larger than that of the pore fluids. The second is the low repeatability of time-lapse survey data due to high amplitude surface-related noise which varies temporally. However, seismic monitoring with FWI of time-lapse borehole seismic data may offer a solution for these issues. FWI is capable of detecting small velocity changes such as those associated with pore fluid substitution. Furthermore, borehole seismic surveys may provide more highly repeatable, higher quality data compared to surface seismic surveys because borehole seismic data is less affected by surface-related noise. This study consists of two parts, a field data analysis and a synthetic study. In the field data analysis, we studied the resolution and repeatability of FWI results at field-data quality, including the presence of actual noise using time-lapse VSP data. VSP data was acquired at the very early stage of EOR and there was no CO2 injection in the time between the two time-lapse VSP surveys. As a result, a high-resolution P-wave velocity model, consistent with a sonic log, was obtained. The P-wave velocity model also revealed excellent repeatability between the two survey data sets. In the synthetic study, time-lapse FWI was performed using synthetic VSP data representing pre- and post- CO2 injection periods. The results of the synthetic study showed that even in the presence of realistic 4D noise, which was estimated in the field data analysis, FWI successfully delineated the distribution of velocity changes caused by CO2 injection when the cross-sectional area of the injection-induced velocity changes were larger than the resolution of the FWI results. With these results, we demonstrated that FWI using time-lapse VSP data was applicable for CO2-EOR monitoring in the field as long as the criteria were met. This conclusion encourages the application of FWI using time-lapse VSP data for CO2-EOR monitoring in onshore Abu Dhabi.