GPU-Based Rendering of Arbitrarily Complex Cutting Surfaces for Black Oil Reservoir Models
BERNARDO BIANCHI FRANCESCHIN, Frederico Abraham, Luiz Felipe Netto, Waldemar Celes · 2019
Numerical simulation of black oil reservoir models is extensively used by the oil industry to predict and plan field exploration. Such simulations produce a large amount of volume data that need to be inspected. One popular visualization technique to inspect volume data is the rendering of cutting surfaces, shaded by mapping properties associated with model elements. In this work, an efficient GPU-based algorithm for rendering arbitrarily complex cutting surfaces for reservoir models is presented. The rendering strategy is based on an efficient point location algorithm. The proposal includes a compact representation of reservoir models in the GPU memory, the use of a compact regular grid as the acceleration technique, and an accurate point location algorithm for handling hexahedral elements with non-planar faces. Computational experiments have demonstrated the effectiveness and efficiency of the proposed approach, even when applied to large models. A set of applications is discussed in the context of inspecting reservoir simulation results.