Enhancing Production Efficiency: AI-Powered Hydrate Detection and Prevention in North Kuwait Jurassic Gas Field

Abstract The Kuwait Oil Company (KOC) and SLB's joint Kuwait Integrated Digital Field (KwIDF) Project for the North Kuwait Jurassic Gas (NKJG) asset is a key initiative in Kuwait's gas production, aiming to boost operational efficiency and productivity via a fully integrated platform. This platform streamlines planning and execution with advanced workflows and automation, connecting multiple disciplines and stakeholders. This integration plays a pivotal role in achieving KOC's ambitious goal of reaching a production capacity of 950 MMSCFD by 2026. The quest to meet production targets demands the optimal use of resources, a task complicated by numerous obstacles. Among these, the formation of hydrates in flowlines and separators resulted in substantial deferments of gas and oil production, according to official reports. To combat this issue, a comprehensive hybrid approach has been adopted, combining machine learning (ML) with a physics-based risk prediction model. The ML model utilizes high-frequency sensor data from the SCADA system to detect hydrate formation events, which are then compared to predictions made by the physics-based model to refine and enhance the accuracy of hydrate risk assessments. This innovative workflow employs compositional fluid modeling software, a tool proven in hydrate prediction and mitigation, to accurately predict hydrate formation temperatures and the effects of methanol concentration on inhibiting the formation of hydrates. The project's methodological approach includes performing thermo-hydraulic simulations to identify potential hydrate formation risks along specific flow paths by calculating the Maximum Hydrate Subcooling Delta Temperature (MHST). This dual approach allows for the real-time detection of hydrate formations and the immediate notification of field operations for swift mitigation actions. The ongoing refinement of the model through the correlation of detected incidents and predicted risks aims to improve the system's accuracy and reliability. The results of this initiative have been promising, with the system operating around the clock to provide early warnings and enabling the field team to take preventative actions promptly. The goal is to achieve over 90% accuracy in model detections, with alarm notifications being issued in under five minutes and providing real-time access for the field operation team. The expected outcome is the elimination of production delays caused by hydrates, potentially recovering 5% of annual gas and oil production, along with cost savings from minimizing chemical injection and hydrate plug removal operations, translating to an estimated annual revenue loss avoidance of $3 million. This project advances the industry by using AI and real-time data to improve predictive maintenance and operational efficiency in dealing with hydrate challenges. Its adaptability suggests it could offer long-term solutions for production delays due to hydrates in gas fields, while also providing valuable insights for the broader industry.