Real-Time Railway Hazard Detection Using Distributed Acoustic Sensing and Hybrid Ensemble Learning

Area of Science:

• Geosciences
• Railway Engineering
• Artificial Intelligence

Background:

• Rockfalls and landslides pose significant geohazards to railway operational safety, particularly on lines like Karabük-Yenice, which traverses mountainous terrain and experiences heavy rainfall.
• Undetected environmental events can lead to severe disruptions and safety risks in railway operations.
• Proactive risk mitigation strategies are essential for ensuring the continuous and safe operation of railways.

Purpose of the Study:

• To develop and validate a real-time system for early detection of environmental phenomena impacting railway lines.
• To integrate Distributed Acoustic Sensing (DAS) with a hybrid ensemble learning model for vibration analysis.
• To enhance railway operational safety by providing timely alerts for geohazards.

Main Methods:

• A 6 km railway corridor in Karabük, Türkiye, was monitored using fiber optic cables and a Luna OBR-4600 interrogator to capture environmental vibrations.
• A hybrid ensemble learning model, specifically a Voting Classifier, was developed using Support Vector Machine (SVM), Random Forest (RF), XGBoost, and Gradient Boosting Classifier (GBC) algorithms.
• Various machine learning algorithms including CatBoosting, SVM, LightGBM, Decision Tree, XGBoost, RF, and GBC were evaluated for signal detection.

Main Results:

• The hybrid Voting Classifier model demonstrated high effectiveness in detecting and classifying environmental disturbances.
• The system achieved 98% accuracy, precision, recall, and F1 score in identifying events such as rockfalls, landslides, and falling trees.
• The study confirmed the capability of fiber optic cables to detect high-intensity vibrations caused by major natural disasters.

Conclusions:

• The integrated DAS and hybrid ensemble learning system provides a reliable method for real-time geohazard detection on railway lines.
• The developed system significantly enhances railway operational safety by enabling early warnings of environmental threats.
• This approach offers a proactive solution for mitigating risks associated with natural disasters in railway corridors.