Verticalization of Sonic Logs Acquired in a Highly Deviated Well from Offshore Abu Dhabi

Abstract

Analysis of single-well sonic measurements to determine formation TI elastic parameters has traditionally been done using deterministic models with multiple, simplistic assumptions using only vertical or horizontal wells with flat structural dip. Here a probabilistic Bayesian-type inversion method is shown, which provides a flexible solution for any wellbore orientation solving for all five TI (transversely isotropic) parameters. This inversion is guided by prior information, which may be; core tests, borehole seismic survey results, offset well data or a public database. The approach demonstrated here uses prior information of TI elastic properties to determine a consistent model at each depth. The inversion uses all sonic slownesses, compressional, fast and slow shear, as well as Stoneley shear and density to provide a continuous output of Thomsen's parameters and anisotropic mechanical properties at each depth. The results are consistent with offset well data as well as walk-above VSP velocities obtained in the same well. This technique was applied on a prominent and thick shale formation, which is present throughout UAE. Most of the wells drilled in UAE penetrate through this shale formation which is locally unstable for drilling, and is a prominent seismic reflector which directly overlies the reservoir. Accurate geomechanical characterization of this shale formation is critical as more wells are drilled at high angles with the application of full 3D finite-element modeling. Similarily the velocity models used for pre-stack depth migration (PSDM) and seismic well tie require accurate anisotropic TI parameters. This workflow is new and applicable to any well orientation or structural dip and yields results that are consistent with offset sonic and borehole seismic measurements. Application of this method will have a profound impact on geophysical velocity and geomechanical model building methods and results.