Curvature attributes with the 3D seismic Kalman filter for fault opening description-An application to a shale oil reservoir in the Jimsar sag, Junggar Basin, Xinjiang

The large-scale shale-oil production in the Jimsar sag of the Junggar Basin is being aggressively developed. To improve unconventional production and make precise suggestions for horizontal well developments, fault interpretations using seis-mic data play an important role. The target reservoir in the Jimsar sag of the Junggar Basin has a comprehensive network of faults, according to the drilled wells and fault identification from the broadband azimuth and high-density 3D seismic data. Several reasons such as the integration between source and reservoir, the substantial abundance of organic matter, and the major reservoir plasticity result in different fault types, which then create different impacts on horizontal wells. We have developed a fault-type identification tool using the curvature attributes via the 3D seismic Kalman filter to provide a more accurate description of fault types. In contrast with previous studies, we derive the Kalman-curvature formulation in terms of space and time, by using a dynamically iterative optimization model. Our strategy reveals that various fault types (open faults, semiopen faults, and sealed faults) represent their characteristics associated with the 3D Kalman-curvature attributes. The result-ant fault divisions are valid by combining the drilled information of lost circulation and fracturing crosstalk of the horizontal wells. High-quality field results validate the robustness of our algorithms.

Automated summary

A fault identification tool using 3D seismic Kalman filter and curvature attributes has been developed to accurately describe fault types in the shale oil region. By applying the Kalman-curvature formulation in space and time, a dynamically iterative optimization model is utilized to reveal the characteristics of various fault types (open, semiopen, and sealed) associated with 3D Kalman-curvature attributes. High-quality field results validate the robustness of this algorithm.