Improving Fault Seal Analysis through Innovative Fault Interpretation

Abstract

Summary Fault seal analyses require valid fault interpretations, and errors in fault interpretation are often attributed to be the cause of unexpected subsurface results. To improve fault interpretation we propose a method that begins with a conservative deep learning technique that provides an incomplete but high confidence fault highlighting visualization. The fault interpretation is completed by combining this initial visualization with a technique that highlights laterally-persistent geophysical anomalies, which include subtle faults near the limit of the resolution of the seismic data. This technique helps to reduce fault interpretation errors, and to quantify the ambiguity of interpreted faults, which can be translated into an uncertainty range in a fault seal analysis.