Drilling Process Monitoring Based on Operation Mode Recognition and Dynamic Feature Extraction

Process monitoring contributes significantly to reducing the risk of downhole faults and preventing undesirable events. This study proposes a process monitoring method based on operation mode recognition and dynamic feature extraction for geological drilling processes. The main idea is to develop different monitoring procedures for various operation modes based on dynamic changes in drilling signals, so as to achieve reliable monitoring for a full drilling cycle, including transient and steady-state processes. The contributions are threefold: an operation mode recognition method is developed for drilling processes based on rules discovered from multivariate time series; a long-short term dynamic feature extraction method is proposed to design a process monitoring method for transient processes; a data-driven model based on the long short-term memory is established for time series prediction to monitor steady-state processes. Industrial case studies from a drilling project demonstrate the effectiveness and superiority of the proposed method.

Automated summary

This study proposes a process monitoring method for geological drilling processes based on operation mode recognition and dynamic feature extraction. It achieves reliable monitoring for the entire drilling cycle, including transient and steady-state processes, by developing different monitoring procedures for various operation modes based on dynamic changes in drilling signals.