AI Cuttings Log Generation via Rig-Offsite Collaboration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for evaluating and characterizing subterranean drilling cuttings are largely manual and time-consuming, making it challenging to automate the process effectively due to variations in size, shape, texture, and color of cuttings, as well as image artifacts like shadowing.
Innovation Solution
A collaborative method involving acquiring images of cuttings at a rig site, generating a clustering of lithology types, transferring data to an offsite location for further evaluation, and generating a description and depth log of labeled lithology types, allowing for semi-automated evaluation and characterization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual evaluation of cuttings images is performed, then evaluation accuracy is maintained, but productivity is reduced and turnaround time increases
Solution Approach 1:
An AI-based intermediary system is introduced between the cuttings images and the final evaluation results. The system processes images through multiple stages (preprocessing, feature extraction, classification, post-processing) with human experts intervening only at critical decision points, thereby increasing throughput while maintaining accuracy
Solution Approach 2:
The evaluation process is segmented into distinct modular stages: image acquisition, preprocessing, feature extraction, classification, and post-processing. Each stage can be independently optimized and processed in parallel, significantly increasing overall productivity while allowing specialized handling at each step
2Productivity
If automated evaluation methods are implemented, then productivity increases, but measurement precision deteriorates due to variability in cutting characteristics
Solution Approach 1:
The system dynamically adjusts processing parameters based on the specific characteristics of each cutting image. Feature extraction parameters, classification thresholds, and preprocessing filters are adapted to account for variations in size, shape, texture, and color, maintaining precision across diverse samples while enabling high-volume processing
Solution Approach 2:
The system incorporates feedback loops where classification results are continuously refined based on confidence scores and expert corrections. Low-confidence predictions are flagged for expert review, and expert feedback is used to retrain and improve the AI models, ensuring high accuracy while maintaining automated processing speed
3Loss of time
If fully automated processing is used, then loss of time is reduced, but reliability decreases due to challenges in handling diverse cutting variations
Solution Approach 1:
The system applies partial automation selectively - fully automated processing is used for high-confidence, routine cases while human experts handle complex or ambiguous cases. This hybrid approach achieves near-real-time turnaround for most samples while maintaining high reliability through expert oversight of critical evaluations
Solution Approach 2:
The system performs preliminary processing steps (image preprocessing, feature extraction, initial classification) automatically before human expert review. This preliminary action prepares and filters data in advance, reducing the time experts need to spend on each case while ensuring reliable final evaluations through expert verification of critical results
Data Source
AI summary
A method for generating a depth log of cuttings obtained during a subterranean drilling operation includes acquiring images of the cuttings and labeling the images with a lagged depth at a rig site; generating a clustering of lithology types in the acquired images at a rig the site; transferring the images and the clustering of lithology types from the rig site to an offsite location; evaluating the images and the clustering of lithology types to label each of the lithology types at the offsite location; and generating a description and/or depth log of the labeled lithology types at the offsite location.