Elastic Full Waveform Inversion With a Permanent Seismic Source ACROSS: Towards Hydrocarbon Reservoir Monitoring

Abstract

Abstract Time-lapse analysis plays an essential role in EOR (enhanced oil recovery) or CCS (carbon capture and storage) management. However, conventional time-lapse seismic often cannot capture differential signals from the target interval because of near-surface heterogeneity or poor signal-to-noise (S/N) ratio. We discuss a potential application of a permanent seismic source "ACROSS" (Accurately Controlled and Routinely Operated Signal System) that can continuously excite controlled seismic signals over years. The superior repeatability of ACROSS makes it possible to subtract any waveforms that do not change over time, and hence the temporal changes due to the fluid replacement at the target interval should be enhanced. Also, it is possible to consider reverse time migration or full waveform inversion rather than just the first arrival or P-wave reflection traveltime inversion, since ACROSS precisely controls the source signature. Both vertical and horizontal single forces are reconstructed by clockwise and counter-clockwise rotations of an asymmetric cylindrical mass. In our synthetic study, we assume that two ACROSS sources are installed at a CO2 pilot injection project in Ketzin, Germany. Elastic FWI (full waveform inversion) is tested with the assumption that the medium parameters (i.e., P- and S- wave velocities and density) before the injection are known. To prevent crosstalk caused by interference of P- and S-waves, we apply a wave separation technique by extracting the scalar (P-wave component) and vector potential (S-wave component) of the elastic wavefields. The simulation results demonstrate that this approach clearly delineates the P-wave velocity decrease caused by the fluid injection. The high-repeatability of ACROSS enables an application of elastic FWI for residual P-wave velocity, which may bring a breakthrough toward CO2 and hydrocarbon reservoir monitoring. Introduction Source repeatability is one of the most important problems in time-lapse surveys and reservoir monitoring. A permanent seismic source "ACROSS" was first developed for detection of sabtle subsurface changes and hense expected for contribution to earthquake prediction (e.g., Kasahara et. al., 2010). The ACROSS source can continuously produce seismic forces by rotations of asymmetric cylindrical mass with either a horizontal or vertical rotation axis. The superior repeatability of ACROSS enables suppression of any waveforms that do not change over time, thereby enhancing the temporal changes due to property changes caused for example by fluid injection and production. Since ACROSS precisely controls the sweep source signature, we discuss applications of advanced seismic technologies rather than that based on first arrival or P-wave reflection traveltime. Reverse time migration (RTM) is known as a high-end imaging method and is widely used for complicated subsurface imaging (Farmer et al., 2006). It is also possible to delineate "zones of change" using time-differential wavefields with just one or two sources with a linear receiver array (Kasahara et al., 2013). It is desirable to quantitatively estimate velocity changes over time using iterative full waveform inversion (FWI), although it is challenging for ACROSS surveys because of limited number of sources available. The ACROSS sources with a horizontal rotation axis can produce both vertical (P-wave dominant) and horizontal (S-wave dominant) seismic forces reconstructed by clockwise and counter-clockwise rotations (e.g., Kasahara et. al., 2011, Kasahara et. al., 2013). Therefore, it is desirable to delineate temporal changes of both P and S wave velocities using elastic imaging algorithm. Here we show test observation data using two ACROSS sources obtained at the Kashiwazaki test field, Japan. Then we performed synthetic study, assuming that two ACROSS sources are installed at CCS pilot test field in Ketzin, Germany (Martens et al., 2013), to discuss possibility and limitation of elastic imaging algorithms.