The Undigitized Frontier: Automating Shale Shaker Analysis with Computer Vision

Traditional drilling rigs rely on intermittent, analog methods for analyzing drilled rock and mud at shale shakers, a crucial indicator of borehole conditions. This undigitized area, particularly the “bucket test” for cuttings return rate, limits real-time insights into hole cleaning effectiveness and borehole stability. Inadequate hole cleaning can lead to costly issues like sticky or stuck pipe, while cavings signal potential borehole enlargement and instability.

To address this, a novel system deploys camera-as-a-sensor technology with AI-enabled image analysis at each shale shaker. Cameras capture images of falling solids, which are then processed by an edge GPU and a central server. Computer vision identifies and characterizes cuttings and cavings, classifying them by size and shape. This real-time data enables calculations of cuttings return and mass flow rates, cuttings size distribution, and comparisons to theoretical drilled rock volume for assessing hole cleaning. Alerts are generated for cavings, allowing for prompt remedial action.

Field deployments in various offshore regions have validated the system’s accuracy in quantifying cuttings return and detecting cavings earlier than conventional methods. This technology optimizes drilling and clean-up times, improves sweep efficiency, detects cuttings accumulation and the onset of cavings, and helps avoid stuck pipe incidents.