Video Based AI Tools for Safety Enhancement on the Drill Floor

Abstract

Abstract This paper explores the utilization of video-based artificial intelligence (AI) tools for enhancing safety measures on the drill floor in the oil and gas industry. It delves into the application of AI-powered systems in monitoring and analyzing critical activities, identifying potential risks, and preventing hazardous incidents. The study showcases the development and implementation of advanced AI algorithms integrated with video monitoring technology, highlighting their effectiveness in real-time risk detection and mitigation. Results demonstrate significant improvements in safety protocols and incident prevention, thereby emphasizing the pivotal role of video-based AI tools in ensuring a safer working environment on the drill floor. The implementation of AI solutions has a profound impact on safety Key Performance Indicators (KPIs). These solutions act as safety tools, capable of accelerating drill performance and fostering a secure work environment on the rig floor. They directly influence the efficiency of drill floor operators, including drillers, assistant drillers, tool pushers, and rig men, while simultaneously reducing the potential for incidents. Based on field data, we can imagine three major challenges during drilling operations: The Lack of Personal Protective Equipment (PPE), Red Zone Management and Latch Monitoring. To address these challenges, high-resolution cameras deliver intricate images of operations on the drill floor. This rich dataset is curated to distinguish between various elements, including various machinery on the drill floor, personnel, and Personal Protective Equipment (PPE). These annotations play a pivotal role in training deep learning algorithms. Incorporating real-world operational data enables the algorithms to grasp the context of each task, bolstering their robustness and accuracy. The techniques prominently applied include segmentation, classification, object detection, pose estimation, and tracking. Owing to the multifaceted image analyses required in real-time, we employ servers equipped with powerful GPUs. All software functionalities run on the edge/rigs, eliminating the need for an internet connection. The real time derived insights are showcased on a web platform, presented as alerts, reports, and dashboards. For swift responses in critical scenarios audible alarms are activated. This ensures immediate interventions to circumvent potential mishaps. Periodic reports are also available, aiding in the refinement of procedures and enhancing the training regimen of the operational teams. Implementation of these solutions yielded key findings: Hazard Detection, Improved Safety, Cost Savings, and Future Potential. This study establishes that Video-Based AI Tools for Safety Enhancement on the Drill Floor represents a paradigm shift in safety management within the oil and gas industry. The findings underscore the significant contribution of these tools in preventing accidents, protecting personnel, and optimizing operational performance. As the industry continues to evolve, the integration of advanced AI technologies stands out as a pivotal strategy to ensure a safer and more sustainable future for drill floor operations.