Full Field Implementation of Physics-Based Analytical Model to Optimize Steamflood in Mukhaizna Heavy Oil Field in Oman

Abstract To optimize steamflood in Mukhaizna Field, Occidental developed connections-based simulation models that retains most of the steam flooding physics. These models address the issue of lengthy run time of traditional 3D reservoir models and resolve the predictability issues of fast physics-less AI/ML models. As these models are being used for full field optimization since 2020, ensuring models reliability and accuracy was a key consideration to optimally operate such large asset. To gain more trust in these models, models input/output, historical match, applied constraints and recommended steam changes are validated with Reservoir Management Teams (RMTs) before field implementation. After implementation, actual steamflood performance is monitored on regular basis with the RMTs relative to model expectations considering assumed constrains. Predictability of the models were checked in the forms of blind tests and comparison of models prediction with actual field data on well level. Furthermore, to QC the reliability of these models, their outputs were validated with actual measurements and reservoir performance for the applicability be used in understanding reservoir dynamics in areas with no/limited data acquisitions. With multiple optimizations and implementations, it was observed that these physics-based models managed to align steam allocation with available opportunities and constraints given the continuous improvement and dynamic challenges in the field. This resulted in observed increase in oil production rate in some areas, reduction in production decline rate in other areas, and increased overall steam efficiency. It was observed also that good handling of production constraints and limitations are essential factors in deciding steam distribution even if reservoir quality and flood efficiency are on the high side. Finally, having history matched physics-based models, enabled the use of model outputs in the day-to-day surveillance, understand subsurface wells connections, unlock hidden opportunities, and provided a reliable tool for optimum operating scenarios going forward. This paper describes one of the largest full field application of physics-based closed-loop reservoir management system in steamflood projects, demonstrate the steps to ensure models reliability, present observed results and challenges, and share lessons learned.