Lightweight Machine Learning and Statistical Models for Robust Fault Detection in IIoT-Driven Oil and Gas Operations
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, cilt.38, sa.10, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 38 Sayı: 10
- Basım Tarihi: 2026
- Doi Numarası: 10.1002/cpe.70757
- Dergi Adı: CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Applied Science & Technology Source, Compendex, INSPEC, zbMATH, Technology Collection (ProQuest)
- Kayseri Üniversitesi Adresli: Evet
Özet
Anomalies in oil and gas systems refer to abnormal changes in the system's operation, including sudden pressure or temperature changes, flow instabilities, or valve malfunctions that interrupt the normal course of operation. These anomalies are important to detect at the initial stage, as they can result in serious consequences, including equipment damage, production process downtime, safety risks, and costly environmental accidents. The introduction of modern oil fields that serve voluminous multivariate sensor information through the IIoT infrastructure makes manual monitoring impractical, and smart automated detection methods should be employed. This paper presents a complex model of anomaly detection in the oil and gas industry based on feature engineering and a blend of machine learning and statistical learning models. Seven Machine Learning algorithms (XGBoost, Random Forest, Support Vector Classifier, K-Nearest Neighbors, Decision Tree, Na & iuml;ve Bayes, and Logistic Regression) and three statistical learning models (Lasso, Ridge, and ElasticNet Logistic Regression) were employed using the publicly available 3W Dataset 2.0.0. To estimate the model robustness, the model was tested under three experiment conditions, clean data, noisy data, and noisy data with hyperparameter optimization using Optuna. The results indicate that the ensemble approaches, especially XGBoost and RF, demonstrated better accuracy and recall in all conditions, with optimized XGBoost to reach 99.93% accuracy and 99.87% recall.