Fault characterization in a postsalt reservoir interval, Jubarte Field (Campos Basin), using seismic attributes and machine learning

Abstract

Seismic attributes are routinely applied to interpretation tasks to improve the recognition of geologic features in the subsurface. Subtle changes in the amplitude and phase components of the seismic data can reveal details and provide information about the management of hydrocarbon reservoirs. To this end, we investigate how instantaneous and geometric attributes can help in fault and discontinuity recognition. Seismic data from Jubarte’s 4D/4C Permanent Reservoir Monitoring system, offshore Brazil, were used to test the impact of a variety of algorithms and parameters in improving fault visualization. The presence of seismic noise and geologic units with weak acoustic impedance contrasts motivated the application of additional methods such as data conditioning workflows and unsupervised machine learning methods. We found that spectral balancing and structure-oriented filtering increased the lateral continuity of some stratigraphic reflectors and attenuated the random noise, which improved fault surface visibility. Attributes were calculated on the full-stack data and azimuth-restricted volumes revealing different impacts of the seismic noise and changes in the lateral continuity of the features. In some cases, faults are better delineated when the acquisition direction is oriented perpendicular (or nearly) to the discontinuity. In addition, the most-positive and most-negative curvature components indicate more details of major features. A large number of seismic cubes and attributes motivated the use of principal component analysis and self-organizing maps, which complement the identification of fault segments with clusters composed of specific neurons aligned within structural discontinuities. The improvements obtained in terms of fault visualization demonstrate the importance of having a workflow that combines different fault and fracture identification methods. An integrated study is important as it identifies effects related to seismic noise and stratigraphic features. For the Jubarte Field, a multiattribute approach demonstrates advantages for delimiting the lateral extension of faults and a more precise discontinuity location.