Integration of deep-learning fault segmentation, horizontal transverse isotropy analysis, and ambient microseismic methods to enhance fracture prediction in the Crisol Anticline, Colombia

Abstract

This study aimed to optimize hydrocarbon production from the naturally fractured reservoirs in the VMM-1 gas field by identifying and interpreting the fault and fracture systems. To achieve this, deep learning fault segmentation was integrated with HTI analysis and ambient microseismic recording. The fault pattern was studied using deep learning fault segmentation, while HTI analysis highlighted the magnitude and distribution of fractures. Ambient microseismic recording was used to identify active faults and fractures. By integrating these three methods, we were able to understand the direction, density, and effectiveness of the various fracture systems, as well as the lateral extent and continuity of the Rosa Blanca Formation. This integration of methods was essential in maximizing ultimate recovery and economic success and has potential applications in the development of other naturally fractured reservoirs.