Project Overview
Industrial facilities face persistent challenges in safety monitoring and operational optimization due to the complexity of real-time data integration and the difficulty of predicting hazardous scenarios before they occur. We develop facility-scale digital twins that synchronize live camera and sensor data with high-fidelity 3D site models, creating an intelligent monitoring system that analyzes operations, tracks personnel and equipment, and predicts potential safety hazards. Our approach combines computer vision, machine learning, and Unity 3D simulation to create a practical feedback loop between perception, simulation, and policy implementation. The technical innovation lies in developing AI-powered models that can identify anomalies, assess risks, and validate safety procedures virtually while maintaining real-time synchronization with physical operations across complex industrial environments.
This research addresses critical safety and efficiency challenges in high-risk industries, particularly oil and gas operations where incidents can have catastrophic consequences. Our digital twin technology enables continuous safety surveillance, automated anomaly detection for pipeline and drilling operations, and simulation-based emergency response training that reduces both accident rates and operational downtime. The system provides predictive maintenance scheduling, compliance monitoring, and root cause analysis capabilities that transform how industrial sites approach risk management and operational optimization. Led by Professor Kenji Shimada and graduate researcher Zee Almusa, our team collaborates with major oil and gas companies and safety organizations to deploy this technology in real-world facilities, establishing new standards for proactive hazard prevention and intelligent industrial monitoring.