SUMMARY We present the advantages of the SeisSol simulation software to model seismic downhole sources and wave propagation in and around the wellbore. The solution algorithm is based on a numerical scheme using the Discontinuous Galerkin (DG) Finite Element approach to solve velocity-stress formulation of the wave equation. The unique property of the numerical scheme is its high approximation order in space and time using fully unstructured tetrahedral meshes to account for geometrically complex computational domains as typically encountered in realistic reservoir applications. Furthermore, local time stepping and space-dependent approximation orders can be applied to reduce computational cost. In this work we take advantage of the flexible mesh refinement strategy of the DG modeling approach to study how borehole-guided waves affect the amplitudes of seismic reflections that occur from small faults or from the boundaries of the reservoir. Furthermore, we investigate how different sonic sources, such as monopole, dipole, or quadrupole sources, can provide different insights in understanding the impact of borehole-guided waves on reservoir formation reflections.
INTRODUCTION In many reservoir development situations, detailed imaging of small faults and micro-fractures is of significant importance to define strategies for maximizing ultimate oil production (Yamamoto et al., 1999). Sonic imaging takes advantage of the higher frequency contents (typically 1-10 kHz) of its downhole sources to image small faults and micro-fractures ahead and around the wellbore trajectory that would be missed by conventional seismic imaging with frequency bandwidth of 2- 500Hz (e.g. Haldorsen et al., 2006). In general, sonic logging is accepted as an invaluable method for obtaining information about the rock formation around the wellbore. However, conventional sonic logging only provides the velocity of the first arrival (P wave). To increase the amount of useful information about the surrounding rock it is necessary to record the full wave train on a dense line of downhole receivers. In order to include P and S waves in our study we will assume a hard rock formation, in particular, as no S waves are recorded in a soft formation and with a conventional monopole sonic tools because there are no critically refracted S waves when the S-wave velocity of the rock is less than the acoustic velocity of the borehole fluid (Chen, 1988). Furthermore, conventional symmetrical monopole sources produce a wave train consisting of a refracted P wave, a refracted S wave, and guided waves including a number of normal modes, Stoneley and Scholte waves. While formation body wave arrivals typically are relatively unaffected by the presence of a steel casing, they may, however, be hard to identify if cement velocities are close to or larger than those of the formation (Tubman et al., 1984). In contrast, the arrivals of the Stoneley wave are strongly influenced by the casing parameters. They respond to the combined effects of the steel, the cement, and the formation. Sinha et al. (2009) study modal propagation of higher-order modes in cylindrical structures and investigates the effect of interaction of steel pipe dispersion with that of the formation.