基于迎风型有限体积法的非结构化网格流场格心数据高质量体绘制方法

High-quality Volume Rendering of Unstructured-grid Cell-centered Data of Flow Fields Based on Upwind FVM

在计算流体力学领域,对复杂流场进行数值模拟时常用非结构化网格划分空间,并广泛采用迎风型有限体积法求解流场。该方法将数值解存储在非结构化网格单元(控制体)中心,产生格心数据,并假设物理量在相邻单元邻接面上形成"间断"。然而,在可视化此类数据时,目前的重构方法破坏了间断约束,导致"流间断"信息丢失,严重制约了绘制质量。为此,在多遍光线投射算法框架下,设计了基于双控制体的采样点间断态重构方法,确保了邻接面上产生间断的流场值,使重构结果如实反映数值计算结果;并针对该重构过程,设计了基于间断片段的图像合成方案,将数值解中的"流间断"信息传递给绘制数据。实验表明,方法能显著提高绘制质量,清晰准确地显示流场内部特征。

In the field of computational fluid dynamics,the upwind finite volume method（FVM） is widely applied to solve the complicated flows with unstructured grids as the volumetric representation.It produces unstructured-grid cell-centered data at the center of each cell（control volume）with the flow discontinuity constraint on the inner-face between face-adjacencies.For visualization,existing reconstruction approaches violate the discontinuity constraint.As a result,the image has low quality due to the lost information of the discontinuity.This research presented a method to reconstruct the discontinuous states at a sample based on the dual control volumes on the framework of multi-pass raycasting.It guarantees that the field on the inner-face is discontinuous between face-adjacencies,which is in accord with the solution of the numerical computation.We also designed a synthesizer for the discontinuous segments based on the reconstruction above,so that the discontinuity information can be transferred to rendering data.The experiments demonstrate that our approach can obviously improve the image quality and clearly visualize the features in the flows.