Abstract With the widespread adoption of digital technologies in the Oil & Gas industry, a range of solutions have emerged to enhance monitoring and deliver actionable insights. While these tools have improved decision-making and efficiency, their growing complexity has made implementation, integration, and scalability increasingly challenging across different operational environments. This paper introduces a data-driven approach to production optimization through the deployment of dynamic well operating envelopes. The objective of this approach is to convert real-time data into operational actions that improve efficiency and minimize downtime. A key innovation lies in the integration of contextual awareness - achieved by using AI to interpret sensor data alongside operational logs. This enables the system to incorporate past events and production constraints into its decision logic, ensuring recommendations are both accurate and field relevant. This methodology integrates data from multiple sources related to well performance - including subsurface measurements, integrity parameters, and production logs - to train machine learning models that associate flow rates with their corresponding operating envelopes and events. These envelopes are then continuously updated using real-time field data, enabling the system to adapt to evolving operating conditions and recommend optimized operational guidelines. The system is structured around a three-step approach: 1. Integration: Aggregates dispersed data sources into a unified digital platform, enabling contextualization of data from individual wells to entire assets. 2. Analysis: Applies advanced analytics and machine learning to identify patterns, detect deviations, and propose optimized guidelines. 3. Recommendation: Generates automated, context-aware recommendations via dynamic operating envelopes, supporting cross-functional collaboration and enhanced asset understanding. The approach was implemented in a mature offshore asset with aging infrastructure and increasing operational constraints. In its initial deployment, the system identified suboptimal operating zones - such as deviations in wellhead pressure and inefficient gas lift configurations - and recommended corrective actions, which were validated and implemented by field engineers. The adaptive nature of the envelopes allowed for real-time tuning in response to evolving reservoir and surface conditions. This responsiveness led to a direct improvement in production efficiency, resulting in a cumulative gain of over 150 kbbl in additional production. The integration of multi-source data - including sensors, logbooks, and operational setpoints - was critical in capturing complex system interactions and enabling prescriptive guidance. These results confirm the potential of dynamic operating envelopes to serve as a foundational layer for autonomous well production optimisation.
Raulino et al. (Mon,) studied this question.