Interactive Simulation and Rendering of SPH-based Fluids

Fluid simulations are very important for Computer Graphics, especially for the special effects industry and the gaming industry. Effect animators use physically plausible simulations to bring life into their scenes. Unfortunatelly, due to their complexity such simulations are time-consuming. However, interactive applications and short production times demand fast simulations and an easy to control setup in combination with an intuitive exploration of the simulation result. The overall goal of this project is therefore to develop physically plausible and interactive fluid simulations which provide an on-the-fly visualization of the simulation result.

Throughout literature many methods exists to simulate fluids. In general they can be summarized in grid-based, so-called Eulerian methods and particle-based, so-called Lagrangian techniques. However, Eulerian methods are accompanied by a strong numerical diffusion which may lead to an unnatural mass-loss and which requires very large grid resolutions. Especially for free surface scenarios Lagrangian methods are much better suited. Therefore, we use well-known SPH methods in combination with prediction-correction steps in order to simulate incompressible fluids. In this context, we focus on fully a GPU-based implementation to make use of the parallel nature of particle systems. In detail, we addressed the following research objectives:

Adaptive particle simulation

Infinite Continuous Adaptivity for Incompressible SPH

In this paper we introduce a novel method to adaptive incompressible SPH simulations. Instead of using a scheme with a number of fixed particle sizes or levels, our approach allows continuous particle sizes. This {enables} us to define optimal particle masses with respect to, e.g., the distance to the fluid's surface. A required change in mass due to the dynamics of the fluid is properly and stably handled by our scheme of mass redistribution. This includes temporally smooth changes in particle masses as well as sudden mass variations in regions of high flow dynamics. Our approach guarantees low spatial variations in particle size, which is a core property in order to achieve large adaptivity ratios for incompressible fluid simulations. Conceptually, our approach allows for infinite continuous adaptivity, practically we achieved adaptivity ratios up to 5 orders of magnitude, while still being mass preserving and numerically stable, yielding unprecedented vivid surface detail at comparably low computational cost and moderate particle counts.

Paper Video The definitive Version was published in ACM Transactions on Graphics, https://doi.org/10.1145/3072959.3073713.

Simulation of thermodynamic processes

Evaporation and Condensation of SPH-based Fluids

In this paper we present a method to simulate evaporation and condensation of liquids. Therefore, both the air and liquid phases have to be simulated. We use, as a carrier of vapor, a coarse grid for the air phase and mass-preservingly couple it to an SPH-based liquid and rigid body simulation. Since condensation only takes place on rigid surfaces, it is captured using textures that carry water to achieve high surface detail. The textures can exchange water with the air phase and are used to generate new particles due to condensation effects yielding a full two-way coupling of air phase and liquid. In order to allow gradual evaporation and condensation processes, liquid particles can take on variable sizes. Our proposed improved implicit surface definition is able to render dynamic contact angles for moving droplets yielding highly detailed fluid rendering.

Paper Video The definitive Version was published in Proc. SCA' 17, https://doi.org/10.1145/3099564.3099580.

Memory efficient neighbor lists for GPU-based SPH simulations

Constrained Neighbor Lists for SPH-based Fluid Simulations

In this paper we present a new approach to create neighbor lists with strict memory bounds for incompressible Smoothed Particle Hydrodynamics (SPH) simulations. Our proposed approach is based on a novel efficient predictive-corrective algorithm that locally adjusts particle support radii in order to yield neighborhoods of a user-defined maximum size. Due to the improved estimation of the initial support radius, our algorithm is able to efficiently calculate neighborhoods in a single iteration in almost any situation. We compare our neighbor list algorithm to previous approaches and show that our proposed approach can handle larger particle numbers on a single GPU due to its strict guarantees and is able to simulate more particles in real time due to its benefits in regard to performance. Additionally we demonstrate the versatility and stability of our approach in several different scenarios, for example multi-scale simulations and with different kernel functions.

Paper   Video

Efficient SPH-based animation of convective-diffusive flows, surfactant dynamics and wetting processes

Fast simulations of incompressible fluids in combination with a diffusive mass transport depend on many parameters: Interactive SPH utilizes small smoothing kernels, adaptive integration time-steps and adaptive sampling of particles. The overall goal is to maintain a consistent simulation while adapting particle resolutions. However, abruptly replacing particles introduces large errors in the density field. Therefore, our temporal blending between particle sets smooths out errors which are introduced into particle fields. We have transferred the latest approaches in the field of incompressible SPH to the GPU in order to take full advantage of the parallel nature of particle simulations.

Temporal Blending for Adaptive SPH

In this paper we introduce a fast and consistent Smoothed Particle Hydrodynamics (SPH) technique which is suitable for convection-diffusion simulations of incompressible fluids. We apply our temporal blending technique to reduce the number of particles in the simulation while smoothly changing quantity fields. Our approach greatly reduces the error introduced in the pressure term when changing particle configurations. Compared to other methods, this enables larger integration time-steps in the transition phase. Our implementation is fully GPU-based in order to take advantage of the parallel nature of particle simulations.

Paper Video The definitive version is available at www.wileyonlinelibrary.com

Beside convective and diffusive transport in the fluid body, interface forces play an important role for natural phenomena, such as wetting effects, surface tension effects and drag/friction effects. However, the strength of surface tension as well as of wetting resistance may depend strongly on the amount of surface active agents (Surfactants) in the fluid. The concentration of surfactants in the transition band between two fluids is mainly controlled by diffusion processes in the fluid body and by adsorption processes at the interface layer. We therefore implement dynamic surface phenomena via convective-diffusive surfactant transport in the fluid-body and by including adsorption/desorption of surfactants at the fluid interface.

Consistent Surface Model for SPH-based Fluid Transport

Surface effects play an essential role in fluid simulations. A vast number of dynamics including wetting of surfaces, cleansing, and foam dynamics are based on surface-surface and surface-bulk in- teractions, which in turn rely on a robust surface computation. In this paper we introduce a conservative Lagrangian formulation of surface effects based upon incompressible smoothed particle hy- drodynamics (SPH). The key concept of our approach is to realize an implicit definition of the fluid’s (free) surface by assigning each particle a value estimating its surface area. Based on this consistent surface representation, a conservative coupling of bulk and surface is achieved. We demonstrate the applicability and robustness of our approach for several types of surface-relevant effects including adsorption, diffusion and reaction kinetics.

Paper  Video

On-the-fly visualization of SPH-based particle fields

In order to speed up volume rendering, we apply adaptive sampling of the volume during ray casting. The degree of adaptivity is controlled by a sampling error analysis framework that is able to derive strict screen space error bounds due to adaptive sampling.

Adaptive Sampling for on-the-fly Rendering of Particle-Based Fluids

We present a fast and accurate ray casting technique for unstructured and dynamic particle sets. Our technique focuses on efficient, high quality volume rendering of fluids for computer animation and scientific applications. Our novel adaptive sampling scheme allows to locally adjust sampling rates both along rays and in lateral direction and is driven by a user-controlled screen space error tolerance. In order to determine appropriate local sampling rates, we propose a sampling error analysis framework based on hierarchical interval arithmetic. We show that our approach leads to significant rendering speed-ups with controllable screen space errors. Efficient particle access is achieved using a sparse view-aligned grid which is constructed on-the-fly without any pre-processing.

Paper  Video

With todays graphics hardware interactive simulation of up to a million of particles are more and more common. The exploration of these particle fields should therefore be applied on-the-fly in order to give a fast feedback of the simulation results. For this purpouse volume renderers utilize the well known emission-adsorption model in order to integrate scalar values along viewing rays. In this context, we propose several caching mechanisms for SPH-based volume ray casting:

Topology-Caching for Dynamic Particle Volume Raycasting

In this paper we present a volume rendering technique for the ad-hoc visualization of interactive particle systems. We focus on methods for an efficient spatial caching (topology caching) of particles when applying a raycasting approach. Thus, we get a fast reconstruction of the scalar field which is defined by the particles' entities. The node-cache allows for efficient caching and pre-fetching of a subset of the octree nodes. The influence-cache provides fast access to all particles which contribute to a specific node including level-of-detail particles. Finally, the introduced slab-cache allows for efficient volume rendering and gradient computation. Our algorithms are completely built and managed on the GPU and interactive frame rates for up to several 100k particles are achieved.

Paper Video Präsentation  Bibtex

Implementation of a GPU-based framework for both simulation and visualization

Fast Single Instruction Multiple Data (SIMD) architectures, for example current graphic card generations, enable a fast parallel computation of SPH physics. Due to their flexible memory handling and generic programming models, current compute APIs, such as CUDA and OpenCL, are well suited for a parallel computation of physics. Most applications also utilize rendering functionality as provided by graphics APIs. This results in a heterogenous system which needs an interoperability between computing contexts. For this purpose we propose a mechanism to integrate compute APIs into todays scene graphs. Please see also our project homepage for more information.

Integrating GPGPU Functionality into Scene Graphs

The concept of scene graphs is widely used in computer graphics to structure graphics-related entities, e.g. geometry, visual attributes as well as abstract data related to certain application requirements like object identifiers or manufacturing details. This paper presents a new method to incorporate General Purpose Graphics Programming Unit (GPGPU)-functionality into scene graph APIs. We define specific scene graph nodes in order to realize a flexible integration of GPU functionality at various levels of granularity without violating the programming paradigm inherent to scene graphs. We focus on current and upcoming compute APIs like CUDA, which are designed for GPGPU purposes. We further present the osgCompute framework that implements our concept on the basis of the OpenSceneGraph API. CUDA is integrated into osgCompute via osgCuda. Our method is flexible in the sense that other compute APIs could be used instead. The advantages of our concept and of osgCuda are demonstrated by presenting examples with different processing requirements.

Paper Präsentation  Projekt  Bibtex

This work is funded by the Siegener Graduate School Development of Integral Heterosensor Architectures for the n-Dimensional (Bio)chemical Analysis.