Enhancing and Accelerating Reservoir Modelling and History Matching of an Offshore Mature Oilfield with Geological and Production System Complexity Using Data-Driven Approach

Abstract

This paper demonstrates that data-driven approach can address to resolve the big data complexity, enhance reservoir characterization and accelerate history matching process into an offshore mature field with many challenges such as complex reservoir geology with high properties variation, fair correlation between seismic amplitude and reservoir properties, multi-stacked completion of production/injection strings, commingled production system from multiple zones, mechanical leaking issue and sparse production allocation. By these complexities, classical technique of integrated full-field modelling cannot be done easily. In addition, it is difficult to acquire reservoir engineering insights for reservoir characterization. Data-driven reservoir modelling approach is applied to tackle the technical challenges. A workflow is proposed for rapid quality check of big data and integrated with systematic reservoir engineering analyses. In data-driven approach, current understanding of geology and physics is relieved with field measurements data as the foundation of constructing the model. The model is kept at high level and introduce further detail where needed within the bound of uncertainty to achieve history match. Data-driven approach has successfully improved reservoir characterization and provide cost-effective technique to accelerate history match process. The quality of big data (i.e. pressure, correct production allocation of each zone) plays key role in data-driven approach. Through early valuable insights from engineers, it significantly reduced the iterations number for history match purposes between engineers and geologist. In this study, it cuts from commonly in months to be weeks. Consequently, field-level and well-level history match can be satisfactorily achieved with deviation within 5%. The blind testing has been conducted to validate data-driven approach and improve confidence level with the model for further field development opportunities such as waterflood optimization, stimulation, new well placement, effective completion and enhanced oil recovery. The results can be reconciled with geological understanding, which will be very useful and suitable in current oil price situation as a cost-effective technique, especially for mature fields with large data and have complexity in geological and production system. The data-driven approach can be deployed to other neighbour mature fields and can improve the level of confidence to support fast business decisions.