From Byproduct to Business Driver: Real-Time Drill Cuttings Intelligence for Safer, Faster, and Smarter Subsurface Decisions

摘要

Summary This study highlights the transformative value of automated drill cuttings analysis as a real-time geoscience tool, especially when downhole tools like MWD and LWD are limited by cost, failure, or extreme environments. By integrating robotics and elemental analysis techniques—namely XRF and LIBS—the system extracts geochemical and mechanical insights from surface-collected cuttings. The process involves automatic collection, cleaning, and pulverization of cuttings, followed by high-resolution elemental analysis. Outputs include modeled mineralogy, brittleness, TOC, elemental gamma ray (EGR), and mechanical proxies, with quality control through standards and barium contamination checks. Machine learning models, trained using PCA and DFA, classify chemofacies with >90% accuracy. Results from over 864 wells demonstrate broad utility: EGR curves replaced failed MWD in the Haynesville and Eagle Ford; geosteering improved through accurate facies classification; redox proxies mapped TOC-rich intervals; silica spikes identified bit wear zones; clay shifts predicted overpressure hazards; and elemental trends helped map subsalt faults and Jurassic boundaries in the Gulf of Mexico. This workflow significantly enhances subsurface understanding, reduces non-productive time (NPT), and enables data-driven decisions in real-time. Its field-proven success makes it valuable not only in hydrocarbon development but also in carbon storage and other subsurface applications.