Machine Learning Drill Bit Vibration Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In earth drilling, vibrational disfunctions such as stick-slip, backward whirl, and high frequency torsional noise reduce drilling efficiency by causing energy loss and damaging drill bits and drill strings due to friction and other interactions between drilling surfaces and geological formations.
Innovation Solution
A machine learning model is trained to classify and identify vibrational disfunctions in real-time using drill bit displacement and acceleration measurements, allowing for the transmission of disfunction types to the surface for parameter adjustments to mitigate these issues, thereby improving drilling efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional vibration monitoring methods are used, then equipment damage can be detected, but drilling efficiency is reduced due to energy loss and inability to real-time adjust parameters
Solution Approach 1:
The system continuously monitors vibrations during drilling operations and provides real-time feedback about disfunction types (stick-slip, backward whirl, etc.). This enables dynamic adjustment of drilling parameters such as RPM and weight on bit, allowing the system to adapt to changing conditions and maintain optimal drilling efficiency while preventing equipment damage.
Solution Approach 2:
The patent replaces traditional mechanical vibration analysis methods with machine learning-based classification systems. The ML models process vibration signals to identify disfunction types, replacing manual mechanical monitoring with automated intelligent systems that can detect and respond to vibration patterns more effectively, thereby improving both reliability and productivity.
2Reliability
If drilling parameters are adjusted frequently to respond to vibrations, then equipment damage can be prevented, but drilling time increases due to continuous adjustments
Solution Approach 1:
The system performs preliminary classification of vibration types using machine learning models before implementing parameter adjustments. By accurately identifying the specific type of disfunction (e.g., cutting-induced stick-slip vs. friction-induced stick-slip) in advance, the system can select the most appropriate pre-planned parameter adjustment strategy, reducing the need for trial-and-error adjustments and minimizing time loss.
Solution Approach 2:
The system changes drilling parameters (RPM, weight on bit, mud flow rate) based on the classified vibration type and severity. By implementing targeted parameter changes rather than continuous adjustments, the system protects equipment effectively while minimizing the total time spent on parameter modifications. The ML model guides parameter selection to achieve the best protection with minimal disruption to drilling continuity.
3Measurement precision
If complex vibration analysis is performed in real-time, then accurate disfunction identification is achieved, but computational resources and system complexity increase
Solution Approach 1:
The machine learning models are trained in advance using historical vibration data and known disfunction patterns. This preliminary training allows the models to make accurate real-time classifications without requiring complex computational analysis during actual drilling operations. The heavy lifting of pattern recognition and feature extraction is performed during the training phase, leaving the real-time system with simpler, faster classification tasks.
Solution Approach 2:
The system creates a digital model (copy) of vibration patterns and disfunction characteristics through machine learning training. Instead of performing complex real-time analysis, the system compares current vibration signals against the pre-learned patterns and classifications stored in the ML model. This copying approach maintains high measurement precision while significantly reducing real-time computational complexity and system requirements.
Data Source
AI summary
A vibrational disfunction machine learning model trainer trains a vibrational disfunction classifier to identify one or more types of vibrational disfunction, or normal drilling, based on measurements of at least one of displacement, velocity, acceleration, angular displacement, angular velocity, and angular acceleration acquired for the drill bit. The vibrational disfunction machine learning model trainer trains the algorithm based on data sets corresponding to characteristic behavior for one or more types of vibrational disfunction and normal drilling. The vibrational disfunction classifier operates in real time, and can operate at the drill bit and communicate vibrational disfunction identification in real time, allowing mitigation of vibrational disfunction through adjustment of drilling parameters.