Towards Well Placement Automation: Revolutionizing Geo-Steering with Automated Dip Picking and Boundary Mapping from Azimuthal Resistivity Multi-Layer Inversions

Abstract Well placement in thin targets and complex geological structures has long been a challenge for successful delivery of horizontal wells. Achieving optimal placement in complex reservoirs requires real-time interpretation of subsurface data and continuous adjustments of the well trajectory. Manual geosteering workflows are subjective, consume time and effort, and are subject to human error. Recent advancements in artificial intelligence and machine learning algorithms utilizing LWD images and multi-layer inversion have enabled the automation of several well placement processes, making a breakthrough in geosteering capabilities and improving consistency, well placement precision, and operational efficiency. Geosteering processes focusing on real-time interpretation and decision-making using Logging While Drilling (LWD) data include conventional LWD curves used for well correlation and petrophysical analysis as well as more advanced data types such as LWD azimuthal images and deep or ultra-deep resistivity inversions. The automated geosteering workflow includes the handling of all these data types and achieves automated lithology evaluation, well-to-well automated correlation, automated dip picking from azimuthal LWD images, automated geological surface updates, as well as automated bed boundary detection derived from advanced deep and ultra-deep resistivity inversions. These developments aim to enhance real-time well placement by streamlining decision-making and reducing dependence on subjective interpretations aimed at ultimately improving operations efficiency. The automated workflows were successfully implemented to correlate between wells, automatically analyze LWD images for dip determination, replacing manual analysis with faster and more consistent interpretation. Correlation algorithms integrated offset well logs to adjust geological surface interpretation while drilling in real time. Boundary mapping automation utilized deep and ultra-deep resistivity data inversions, enabling the precise identification of distance to bed boundaries. An integrated workflow integrates all these components as well as directly feeds directional drilling automation platforms for accurate and immediate trajectory adjustments. The drilling and geosteering automated process collectively results in reduced wellbore placement deviations, improved borehole tortuosity, maximized reservoir contact, and significant time savings. This paper introduces a fully automated drilling and geosteering framework that transforms the traditional real-time decision-making process, which sets a new standard for the accuracy and efficiency of horizontal well delivery. This is an important step towards fully automated geosteering and drilling automation.