Realtime Application of Early Kick and Loss Detection (EKLD) for the Entire Drilling Operation

Abstract In drilling operations, deploying a real-time monitoring and detection model to identify early kick and loss events is essential for ensuring safety and achieving operational success. The model presented has the capability to accurately recognize various phases of the drilling process—such as active drilling, tripping in, reaming, and tripping out—from the available real-time data. Furthermore, the system must detect kick/loss events across all phases accordingly, providing operators with meaningful and reliable results. Using a smart drilling phase detection algorithm in real-time, then applying unique early kick/loss detection according to the drilling activity to achieve EKLD for the entire drilling operation. Additionally, the monitoring system is designed to address the challenge of performing complex calculations in real time, ensuring that there is no lag in the system's responsiveness. The model was validated and evaluated using real-time drilling data directly from the rig in real-time, covering the entirety of drilling operations. Field testing demonstrated its effectiveness in detecting both kick and lost circulation events across all phases of drilling, applicable especially for real-time monitoring. A comprehensive performance comparison with traditional kick detection models was conducted. This comparison shows that the model is more suitable for all phases of drilling activities, by having its unique detection and validation algorithms tailored for each activity. In conclusion, the model stands out for its ability to effectively detect kicks and losses throughout the entire drilling operation, across various drilling phases, as well as its capability to process all data in real time. The model presented in this paper represents one of the first advanced systems capable of processing real-world drilling data for the entire drilling operation. It achieves accurate and timely detection of early kicks and losses in real-time while minimizing false alarms.