基于CUDA的弱可压SPH流体建模与仿真

为了实现小尺度范围流体场景的实时、真实感模拟,采用弱可压SPH方法对水体进行建模,提出了流体计算的CPU GPU混合架构计算方法。针对邻域粒子查找算法影响流体计算效率的问题,采用三维空间网格对整个模拟区域进行均匀网格划分,利用并行前缀求和和并行计数排序实现邻域粒子的查找。最后,采用基于CUDA并行加速的Marching Cubes算法实现流体表面提取,利用环境贴图表现流体的反射和折射效果,实现流体表面着色。实验结果表明,所提出的流体建模和模拟算法能实现小尺度范围流体的实时计算和渲染,绘制出水的波动、翻卷和木块在水中晃动的动态效果,当粒子数达到1 048 576个时,GPU并行计算方法相较CPU方法的加速比为60.7。     

GPU构网的GeoMipMap地形无缝绘制算法

针对传统CPU实时构网算法和预处理阶段构网算法速度较慢问题,提出一种GPU构网的GeoMipMap地形渲染算法.算法的构网阶段由GPU实现,将CPU从繁重的构网工作中解放出来,并大幅度减少CPU向GPU传输的数据量,提高地形的渲染速度.整个地形分成大小相等的若干地形块,每个地形块又分为内部及四条边共五部分,对这五部分按分辨率不同分成多个细节层次,为每个细节层次计算空间误差,渲染时各部分根据屏幕投影误差选择细节层次,所构网格更加符合地形表面特征.考虑到GPU构网算法的高度并行性,采用一种新的裂缝处理方式,四条边的屏幕投影误差以边上顶点的空间误差计算,使得相邻块对于共享边的细节层次的计算结果相同,从而保证相邻块间无裂缝,且网格连续.实验结果表明该方法能够以较高的质量完成大规模地形的实时平滑漫游.     

基于曲线拟合函数和GPU的地形无缝渲染

针对大规模地形实时渲染时不同层次细节过渡产生的画面不连续性问题,以及计算机实时调度大量数据而造成帧速较低的问题,提出了一种基于曲线拟合函数和GPU加速的地形实时绘制方法。根据采样数据点采用最小二乘法构造曲线拟合函数,通过曲线函数控制不同层次网格顶点的布局,从而消除因层次细节变化产生的裂缝。同时根据分辨率不同构建金字塔模型,针对不同层次细节区域数据进行有损或无损压缩,依据视点运动预测实时解压相应数据。实验表明,该方法在地形实时渲染阶段,在保证了较高帧速率的同时,利用视点运动预测保证了帧速率变化小,有效的消除了裂缝,增强了画面效果。     

基于GPU投影网格的曲面渲染技术

研究曲面渲染技术对船舶、汽车、飞机造型虚拟的设计飞行器飞行实时仿真系统设计尢为重要.为了克服传统的曲面渲染方法的不足和提高实时性,充分利用图形处理器(GPU)不断提高的渲染能力,包括GPU的可编程性和高度并行计算特性,在GPU上实现了投影网格(Projected Grid)的视点相关的曲面渲染技术.从视点发出的投射光线穿过投影网格后,将根据可视化细节的重要程度,自动生成具有不同细节层次(Levels of Detail)的曲面网格,并且实时地更新网格的细节层次需求.在整个渲染过程中保持稳定的帧率,生成与视点相关的曲面光滑流畅.试验证明满足了实时交互性的要求,在工程虚拟仿真领域有广泛的应用前景.     

基于硬件细分的层次细节地形渲染算法

针对顶点着色器细分地形网格需要额外生成模板、计算细分层次复杂的不足,提出了一种利用细分着色器进行地形网格细分的层次细节(LOD)地形渲染算法。利用分块四叉树组织建立地形粗糙网格的分层结构,以LOD判别函数对活动地形块进行筛选;提出了在细分控制着色器中基于视点三维连续距离的细分因子计算方法,并针对外部细分因子进行处理消除了裂缝;实现在细分计算着色器上的置换贴图,对精细网格的高度分量进行位移。而且将四叉树结构存储至顶点缓冲区,减少中央处理器(CPU)与图形处理器(GPU)的资源交换;引入细分队列加速细分过程。实验证明,该算法具有平滑的细节层次过渡和良好的细分效果,能够有效提高GPU利用率和地形渲染效率。     

微分域网格变形的GPU加速算法

微分域网格变形方法能够较好的保持网格模型的局部细节特征,但其计算需要耗费较长的时间.结合GPU的高速并行运算性能,设计并实现了一种基于GPU的微分域网格变形算法.通过GPU进行网格的微分坐标求解、线性系统系数矩阵的Cholesky分解、线性系统求解等运算,从而将网格局部细节特征编码和解码过程以及变形结果的绘制完全通过GPU完成.实验结果表明该算法能够有效加速微分域网格变形方法的计算和绘制.     

基于动态网格的云的实时仿真

为实时绘制真实感的云场景,针对已有基于网格的方法效率低下的问题,提出动态网格方法,充分考虑风对云的发展变化的影响.在动态网格上离散化云的运动方程和热动力学方程,根据视距的变化调整网格分辨率.在云的绘制过程中,将Cornette-Shanks相位函数用于多重前向散射模型,并采用GPU加速来提高仿真效率.实验结果表明,提出的动态调整网格分辨率的方法提高了仿真效率,可以实时模拟云与物体的交互和绘制真实感的云图像.     

基于BRDF 和GPU 并行计算的全局光照实时渲染

基于光线追踪,将屏幕图像像素分解为投射光线与场景对象交点面片辐射亮度和 纹理贴图的合成,每个面片的辐射亮度计算基于双向反射分布函数(BRDF)基的线性组合,并通 过图形处理器(GPU)处理核心并行绘制进行加速,最后与并行计算的纹理映射结果进行合成。 提出了一种基于BRDF 和GPU 并行计算的全局光照实时渲染算法,利用GPU 并行加速,在提 高绘制效率的前提下,实现动态交互材质的全局光照实时渲染。重点研究：对象表面对光线的 多次反射用BRDF 基的线性组合来表示,将非线性问题转换为线性问题,从而提高绘制效率; 利用GPU 并行加速,分别计算对象表面光辐射能量和纹理映射及其线性组合,进一步提高计算 效率满足实时绘制需求。     

基于CUDA的并行加速渲染算法

GPU可以快速有效的处理海量数据,因此在近些年成为图形图像数据处理领域的研究热点。针对现有GPU渲染中在处理含有大量相同或相似模型场景时存在资源利用率低下和带宽消耗过大的问题,在原有GPU渲染架构的基础上提出了一种基于CUDA的加速渲染方法。在该方法中,根据现有的GPU渲染模式构建对应的模型,通过模型找出其不足,从而引申出常量内存的概念;然后分析常量内存的特性以及对渲染产生的作用,从而引入基于常量内存控制的方法来实现渲染的加速,整个渲染过程可以通过渲染算法进行控制。实验结果表明,该方法对解决上述问题具有较好的效果,最终实现加速渲染。     

在GPU上实现地形渲染的自适应算法

为了满足飞行状态实时监控系统对地形渲染快速逼真的要求,提出一种基于GPU的交互式地形自适应渲染算法.该算法中每帧渲染包含2遍GPU处理过程:第1遍采用流计算的方式,利用固定网格映射方法生成粗糙地形采样网格,在此基础上,根据地形表面复杂度计算粗糙采样网格中每个三角形的细化深度;第2遍进行地形的渲染,根据第1遍计算出来的每个面片的细化等级,选择初始化阶段预存储在GPU缓存中的不同细化等级的网格模板,对粗糙采样网格进行自适应细分,为了消除T型连接的问题,在顶点着色器中包含了一个网格模板的匹配处理过程.最后采用高程数据配合卫星照片的方式,生成具有高度仿真的三维虚拟地貌.基于文中算法实现的实时监控系统在支线飞机的飞行试验中取得了较好的效果.     

小型水域水面涟漪的模拟方法

以实际项目为研究背景,提出了一种小型水域水面涟漪模拟方法。将Johanson 的投影网格法用于水面建模,采用二维波动方程描述水波运动,运用显式有限差分法和可编程GPU技术实现方程求解;并采用适当的渲染算法,比较真实地模拟了雨滴落入水面后水面波动的现象,在网格为256×256,图像大小为1 024×768 的情况下,刷新频率达到了85 帧以上,满足了实时性的要求。     

面向虚拟环境的真实感漫游湖面建模算法

提出了一种面向虚拟漫游的太湖湖面建模算法。引入了Johanson投影网格法,避免了多分辨率LOD的龟裂问题,将视空间里生成的网格投影到世界空间中。为了提高渲染效率,避免对投影到世界空间中的每个湖面网格点进行噪声计算,采用了准均匀B样条曲面来构造湖水水面,将Perlin噪声施加于B样条曲面的特征控制点,间接地控制水面高度。依靠图形硬件的加速功能保证了B样条曲面的生成速度。实验表明,用该算法可以在普通PC平台上实时模拟漫游太湖的真实场景。     

Woodification: User‐Controlled Cambial Growth Modeling

We present a botanical simulation of secondary (cambial) tree growth coupled to a physical cracking simulation of its bark. Whereas level set growth would use a fixed resolution voxel grid, our system extends the deformable simplicial complex (DSC), supporting new biological growth functions robustly on any surface polygonal mesh with adaptive subdivision, collision detection and topological control. We extend the DSC with temporally coherent texturing, and surface cracking with a user‐controllable biological model coupled to the stresses introduced by the cambial growth model.     

基于笛卡尔切割单元法的复杂河道地理信息系统环评可视化

复杂河道中污染物扩散计算及其在地理信息系统(GIS)上的可视化,对于地表水环境影响评价(EIA)具有非常重要的意义,但在网格生成、污染物计算模型及计算结果可视化方面存在着诸多困难。针对点源岸边排放河流污染物计算及基于GIS可视化中的难点问题,提出了基于切割单元法的地面水环境影响评价可视化方法。将切割单元法应用于网格剖分,通过切割单元交点追踪算法及河道轮廓线内背景网格筛取算法,生成了复杂河道笛卡尔网格。提出了基于污染物二维稳态衰减模式的网格自适应加密与稀疏算法,在非结构化笛卡尔网格基础上采用了基于河流几何信息判断的点源岸边排放河流污染预测算法与区域填充算法,实现了环境影响评价计算结果的可视化显示。通过一个河流污染环境影响评价可视化的实例,验证了所提方法的可行性与有效性。     

基于图形处理器的水面仿真

水面效果的仿真可大幅提高自然环境仿真的真实感,传统对于CPU的仿真存在占用CPU时间和系统资源的缺点,针对存在问题,建立了基于图形处理单元（GPU）的水面仿真方法,讨论水面特效在GPU上的实现、以及水面网格在GPU中的重构。因为运算以及水面网格重构都在GPU中完成,充分利用GPU强大的图形处理能力,因此不会造成额外的系统开支,并且增强了对水面细节的表现,使得水面的逼真度和实时性增强。     

Frame-sliced voxel representation: An accurate and memory-efficient modeling method for workpiece geometry in machining simulation

This paper presents a new and enhanced voxel representation format for modeling the machined workpiece geometry in simulating machining operations involving repeated update of the workpiece model volume. The modeling format is named as the Frame-Sliced Voxel representation (FSV-rep) as it uses a novel concept of frame-sliced voxels to represent the boundary of the workpiece volume. The FSV-rep uses a multi-level surface voxel representation for sparse and memory-efficient implementations. The utilization of frame-sliced voxels enables approximation of the workpiece surface to only loosely depend on the grid resolution but achieve sub-voxel resolution updates for the model volume. It can, thus, provide a boundary representation of the workpiece model at an accuracy that is much higher than a basic voxel model of the same grid resolution and a similar model size. Quantitative comparisons of the FSV-rep with the traditional voxel representations at the same finest grid resolution show improvement up to two orders of magnitude in accuracy with only marginal increases in the model size. This confirms the effectiveness of the FSV-rep in simulating machined workpiece geometry in complex machining processes such as multi-axis milling.     

Stable free surface flows with the lattice Boltzmann method on adaptively coarsened grids

In this paper we will present an algorithm to perform free surface flow simulations with the lattice Boltzmann method on adaptive grids. This reduces the required computational time by more than a factor of three for simulations with large volumes of fluid. To achieve this, the simulation of large fluid regions is performed with coarser grid resolutions. We have developed a set of rules to dynamically adapt the coarse regions to the movement of the free surface, while ensuring the consistency of all grids. Furthermore, the free surface treatment is combined with a Smagorinsky turbulence model and a technique for adaptive time steps to ensure stable simulations. The method is validated by comparing the position of the free surface with an uncoarsened simulation. It yields speedup factors of up to 3.85 for a simulation with a resolution of 480³ cells and three coarser grid levels, and thus enables efficient and stable simulations of free surface flows, e.g. for highly detailed physically based animations of fluids.     

GPU-based high-performance computing for integrated surface–sub-surface flow modeling

The widespread availability of high-resolution lidar data provides an opportunity to capture micro-topographic control on the partitioning and transport of water for incorporation in coupled surface – sub-surface flow modeling. However, large-scale simulations of integrated flow at the lidar data resolution are computationally expensive due to the density of the computational grid and the iterative nature of the algorithms for solving nonlinearity. Here we present a distributed physically based integrated flow model that couples two-dimensional overland flow and three-dimensional variably saturated sub-surface flow on a GPU-based (Graphic Processing Unit) parallel computing architecture. Alternating Direction Implicit (ADI) scheme modified for GPU structure is used for numerical solutions in both models. Boundary condition switching approach is applied to partition potential water fluxes into actual fluxes for the coupling between surface and sub-surface models. The algorithms are verified using five benchmark problems that have been widely adopted in literature. This is followed by a large-scale simulation using lidar data. We demonstrate that the method is computationally efficient and produces physically consistent solutions. This computational efficiency suggests the feasibility of GPU computing for fully distributed, physics-based hydrologic models over large areas.     

A traffic cellular automata model based on road network grids and its spatial and temporal resolution’s influences on simulation

Microscopic Traffic Simulation Model based on Cellular Automata (CA) has become more and more popular since it was firstly introduced by Creamer and Ludwig in 1986. Cellular automata are simpler to implement on computers, provide a simple physical picture of the system and can be easily modified to deal with different aspects of traffic. However, in a traditional traffic CA model, the spatial resolution of CA and temporal resolution of simulation are low. Take TRANSIMS for example. The size of cellular automata is 7.5 m and the time step equals 1 s. In such a case, if a vehicle drives at a speed of 4 cells per s, the speed difference between 95 km/h (3.5 1 7.5 m/s) and 121 km/h (4.4999 1 7.5 m/s) will not be distinguished by simulation models. And the temporal resolution of 1 s makes the system hard to model different drivers’ reaction time, which plays a very important role in vehicular movement models. In this paper, a microscopic traffic cellular automata model based on road network grids is proposed to overcome the low spatial and temporal resolutions of traditional traffic CA models. In our model, spatial resolution can be changed by setting different grid size for lanes and intersections before or during simulation and temporal resolution can be defined according to simulation needs to model different drivers’ reaction time, whereas the vehicular movement models are still traditional CA models. By doing so, the low spatial and temporal resolution of CA model can be overcome and the advantages of using CA to simulate traffic are preserved. The paper also presents analyses of the influences on simulation of different 1D lane grid size, 2D intersection grid size and different combinations of temporal resolution and mean drivers’ reaction time. The analysis results prove the existence of spatial and temporal resolution thresholds in traffic CA models. They also reveal that the size of grids, the combinations of different temporal resolutions and mean drivers’ reaction time do pose influences on the speed of vehicles and lane/intersection occupancy, but do not affect the volume of traffic greatly.     

Assessment of multiscale resolution for hybrid turbulence model in unsteady separated transonic flows

This paper presents results of a computational study conducted to assess the multi-scale resolution capabilities of a hybrid two-equation turbulence model in predicting unsteady separated high speed flows. Numerical solutions are obtained using a third order Roe scheme and the SST (shear-stress-transport) two-equation-based hybrid turbulence model for three-dimensional transonic flow over an open cavity. A detailed assessment of the effects of the computational grid and the hybrid turbulence model coefficient is presented for the unsteady flow field. Computed results are presented for both the resolved and the modeled turbulent kinetic energy (TKE) and for the predicted sound pressure level (SPL) spectra, which are compared to available experimental data and large Eddy simulation (LES) results. The comparison shows that the predicted SPL spectra agree well with the experimental results over a frequency range up to 2500 Hz, and that hybrid turbulence effectively models the shorter wavelengths. The results demonstrate improved agreement with experimental SPL spectra with increased grid resolution and a reduced hybrid turbulence model coefficient. In addition, they show that energy dissipation of the unresolved scales is over-predicted at low resolutions and that the hybrid coefficient influences the grid resolution requirements.