Real-time Sensing and Modeling for Rotary Drilling Operations

This paper presents an integrated real-time sensing and modeling system for rotary drilling operations, utilizing synchronized hydraulic, pneu-matic, and visual sensors. The proposed frame-work combines composite feature engineering, adaptive thresholding, and multi-state classification to enable timely anomaly detection and intelligent operational recommendations. A comprehensive and intuitive user interface delivers real-time visualization of all key sensor values and provides immediate feedback and alerts to operators, thereby supporting data-driven decision-making in the field. Field deployments on mul-tiple construction sites in Hong Kong demonstrate the effectiveness, practicality, and scalability of the system in identifying abnormal drilling events and significantly enhancing operational safety and efficiency. This work lays the foundation for intelligent automation in pile foundation construction, offering a robust and extensi-ble solution for future smart construction applications.