The management of reservoir simulation models has been an important need of engineers in petroleum industry. However, due to data sharing among reservoir simulation models, data replication is common and poses many challenges to model management, including management efficiency and data consistency. In this paper, we propose a data component based methodology to manage reservoir simulation models. It not only improves management efficiency by removing data replicas, but also facilitates information reuse among multiple models. We first identify the underlying structure of the simulation models and decompose them into three types of components: reservoir realization, design, and simulator configuration. Our methodology then identifies the duplicate components and guarantees that each component has one physical copy in the data repository. By separating the logical connections between the models and the components from the physical data files, our methodology provides a clean and ...
Cong Zhang, Amol Bakshi, Viktor K. Prasanna