A New Workflow of Drilling Anomaly Detection Based on Prior Knowledge and Unsupervised Learning
Bibliographic record
Abstract
Summary Timely detection of drilling anomalies is critical for reducing nonproductive time and ensuring operational safety, especially in deep, complex, or high-pressure environments. This study proposes a novel anomaly detection workflow that integrates domain-specific prior knowledge with an unsupervised learning algorithm, affinity propagation (AP), to overcome limitations of conventional rule-based and purely data-driven methods. The framework consists of four core stages: real-time data preprocessing, drilling state identification with parameter selection, fluctuation sensing using the PWDTW-AP model (prior weight dynamic time warping-AP), and anomaly classification through a drilling anomaly index (DAI). Methodologically, the PWDTW-AP model combines three key innovations: (1) an exemplar-based clustering algorithm that avoids manual cluster number specification and allows adaptive sensitivity control, (2) dynamic time warping (DTW) to suppress the impact of regular or periodic fluctuations, and (3) prior weight (PW) to emphasize critical anomaly-sensitive parameters. This synergy enables the model to balance high sensitivity with strong robustness in noisy field conditions. Compared with K-means, long short-term memory (LSTM), and Bayesian delayed rejection adaptive Markov chain Monte Carlo (DRAM-MCMC) models, DTW-AP achieves 100% event-based recall (ER), with only 0.3 average false alarms per event and a fast response time (RT) of 14.6 seconds. Integrating PW improves sensitivity to key parameters while avoiding false alarms in the test cases. In field deployment, the workflow first identifies the current drilling state from streaming data, dynamically selects monitoring parameters, and continuously updates feature weights based on historical cases. Anomalies are detected through parameter fluctuation clustering and further interpreted using the number and transition of clusters alongside DAI, enabling efficient classification and severity analysis. Field tests across multiple well cases confirm its ability to detect multiple types of anomalies with strong interpretability and real-time performance, making it a scalable solution for intelligent drilling anomaly detection.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".