Application of risk-informed closed-loop field development workflow to elucidate the evolution of uncertainties

• Published in: Journal of petroleum science & engineering 2021-02, Vol.197 (C), p.107960, Article 107960
• Main Authors: Loomba, Ashish Kumar, Botechia, Vinicius Eduardo, Schiozer, Denis José
• Format: Article
• Language: English
• Subjects: Data assimilation; Decision-making; Field development; Optimization; Reservoir simulation; Uncertainties
• ISSN: 0920-4105; 1873-4715
• DOI: 10.1016/j.petrol.2020.107960

Closed-loop field development (CLFD) is an exhaustive combination of multidisciplinary tasks to use frequently acquired data for optimizing a pre-defined objective function of the field development plan (FDP). Although new information is bound to decrease the uncertainty around field development, previous studies have shown that CLFD could fail for several theoretical reasons. In this work, a risk-informed CLFD process is introduced to increase the chances of success of the optimized FDP in the true field. A risk-informed CLFD utilizes insights from a systematic approach for evaluating risks associated with field development to make robust decisions for the true field. We implemented the risk-informed CLFD methodology on two different case studies: (I) a scenario with mostly horizontal wells and, (II) a scenario with all vertical wells. While one of the previous studies has shown that CLFD can decrease the NPV by 2% for the presented case study I, our workflow validates the importance of the risk-informed CLFD by improving the net present value (NPV) of the project by 14%. Implementation of CLFD on case study II validates the workflow once again by improving the NPV by 40%. As with previous studies, we considered the project objective function (i.e., NPV) as the key performance indicator of the CLFD. While the performance indicator suffices the requirement of evaluating our methodology, in this work we delved deeper to understand how the intermittently acquired information influences the ensemble of simulation models and uncertainty assessment. We discuss that further fine-tuning the objective function of the optimization problem can improve the likelihood of success in the true field. The paper presents two case studies that are based on a field-scale benchmark model in an attempt to answer the questions about the purport of a field development process with multiple phases while acquiring and utilizing information intermittently. Also, the work validates the risk-informed CLFD methodology to encourage tests on more complex fields. Some key observations to improve the CLFD methodology further are also discussed in the work. •Field development plan can be improved by closed-loop field development.•Recurrently acquired information helps but does not capture uncertainties entirely.•A risk-informed workflow is vital to make risk-averse decisions.•Focus on model accuracy while accounting bias and error can help improve decisions.•Quantification of closed-loop f