Revolutionizing Well Data Access and Analysis with AI: A Case Study and Insights Major Middle Eastern Oil Fields

In the quest for operational excellence, petro-technical experts require comprehensive visibility into the myriad data and events spanning the lifecycle of each well. This task is daunting, given that crucial data and expertise are dispersed across numerous databases, thousands of documents, and among hundreds of experts, many of whom are retired or nearing retirement. To address this challenge, operator has developed an integrated digital knowledge layer for wells, aimed at empowering users to easily access information, events, and knowledge related to operating wells. This facilitates periodic 360-degree well reviews to ensure wells are performing as expected and to identify opportunities for performance enhancement. The AI Digital Well Book, an in-house solution, offers a unified view of diverse data elements across various disciplines. Each well and its associated data elements are organized into chapters, capturing key events, gathered data, analyzed data, and live data, represented as individual pages. AI Digital Well Book is architected around three synergistic layers. Data Integration and Synthesis Layer (Foundation Layer): It includes key functions such as data ingestion, semantic normalization, event extraction and knowledge graph linking entitiesAI and Analytics Layer (Intelligence Layer): This layer applies advanced analytics and AI models to derive insights, detect anomalies, and support decision-making.Interaction layer: This top layer provides an intuitive, conversational and Visualization interface for engineers and domain experts to interact with the system. Digital Well book allows users to visualize all events on a timeline, irrespective of data storage locations. The system extracts events from data, presents them within the well lifecycle, and connects these events through common oil and gas entities and concepts. It organizes and indexes data around well concepts such as equipment, vendors, well activities, problems, maintenance history, intervention events, reservoir monitoring, rig-less performance analysis, drilling issues, reservoir and petroleum engineering analysis, health and safety events, and incidents. The solution supports well and reservoir management review processes in a standardized, structured, and centralized manner, significantly enhancing knowledge sharing between assets by documenting best practices and lessons learned across 5000+ wells. Engineers have experienced a 60% reduction in time spent searching for data during well reviews. Additionally, the solution enables easy and effective searches of historical information, facilitating fact-based, informed decisions regarding well abandonments, interventions, and planning efforts. It provides a single source of data, faster access, improved quality checks, and integrated workflows, thereby simplifying collaboration. A notable feature of this solution is Generative AI real time data agent, an interactive chat-based tool that allows petroleum and reservoir engineers to interact with well data contextually. This agent highlights issues related to wells, enhancing the engineers’ ability to manage and optimize well performance.