Advanced Visualization for Reservoir Simulation

Abstract

Abstract This paper provides an overview of an advanced visualization system for visualizing giga-cell simulation models to assist Reservoir Engineers in managing and developing Saudi Arabia’s giant fields. The challenges in the hydrocarbon industry require the use of the latest technology to maximize recovery in a cost-effective manner. Over the last few years, reservoir simulation activities have undergone major transformation toward the use of giga-cell simulation models [1]. This transformation stems from the use of advanced data acquisition solutions and advanced modeling packages to capture and model physically large fields. Fine scale-grid blocks capture detailed description of geologic heterogeneity that is necessary to model the complexity of fluid flow in the subsurface. Giga-cell simulation models are currently generated using Saudi Aramco’s GigaPOWERS parallel reservoir simulator [2]. The size of reservoir simulation models continues to grow to exceed the billion-cell barrier. The design of a high performance computational platform for simulation of giant reservoir models has been discussed in other literature [3]. In this paper, we focus on the simulation post-processing visualization system developed, which consists of advanced technologies for handling and visualizing massive amounts of data. Implemented techniques include: Remote visualization utilizing graphics processing clusters, parallel data loading, level of detail graphics rendering and hierarchal multi-resolution data structure. In addition, a number of advanced visualization features that have been implemented are highlighted in this paper, such as the ability to display fluid patterns in the field by utilizing streamlines and vector fields, advanced volume visualization techniques based on streamlines and wells filtering to help engineers better understand the fluid flow in subsurface reservoirs. Furthermore, the performance considerations, challenges and impact on reservoir simulation studies are also discussed.