Adam W. Bargteil

Graphical modeling and animation of ductile fracture

This video demonstrates a method for realistically animating ductile fracture in common solid materials such as plastics and metals. The effects that characterize ductile fracture occur due to interactions between yielding plastic and the fracture process. By modeling this interaction, this ductile fracture method can generate realistic motion for a much wider range of materials than could be realized with a purely brittle model.

A method for animating viscoelastic fluids

This paper describes a technique for animating the behavior of viscoelastic fluids, such as mucus, liquid soap, pudding, toothpaste, or clay, that exhibit a combination of both fluid and solid characteristics. The technique builds upon prior Eulerian methods for animating incompressible fluids with free surfaces by including additional elastic terms in the basic Navier-Stokes equations. The elastic terms are computed by integrating and advecting strain-rate throughout the fluid. Transition from elastic resistance to viscous flow is controlled by von Misess yield condition, and subsequent behavior is then governed by a quasi-linear plasticity model.

A semi-Lagrangian contouring method for fluid simulation

In this article, we present a semi-Lagrangian surface tracking method for use with fluid simulations. Our method maintains an explicit polygonal mesh that defines the surface, and an octree data structure that provides both a spatial index for the mesh and a means for efficiently approximating the signed distance to the surface. At each timestep, a new surface is constructed by extracting the zero set of an advected signed-distance function. Semi-Lagrangian backward path tracing is used to advect the signed-distance function. One of the primary advantages of this formulation is that it enables tracking of surface characteristics, such as color or texture coordinates, at negligible additional cost. We include several examples demonstrating that the method can be effectively used as part of a fluid simulation to animate complex and interesting fluid behaviors.

A finite element method for animating large viscoplastic flow

We present an extension to Lagrangian finite element methods to allow for large plastic deformations of solid materials. These behaviors are seen in such everyday materials as shampoo, dough, and clay as well as in fantastic gooey and blobby creatures in special effects scenes. To account for plastic deformation, we explicitly update the linear basis functions defined over the finite elements during each simulation step. When these updates cause the basis functions to become ill-conditioned, we remesh the simulation domain to produce a new high-quality finite-element mesh, taking care to preserve the original boundary. We also introduce an enhanced plasticity model that preserves volume and includes creep and work hardening/softening. We demonstrate our approach with simulations of synthetic objects that squish, dent, and flow. To validate our methods, we compare simulation results to videos of real materials.

Fast simulation of mass-spring systems

We describe a scheme for time integration of mass-spring systems that makes use of a solver based on block coordinate descent. This scheme provides a fast solution for classical linear (Hookean) springs. We express the widely used implicit Euler method as an energy minimization problem and introduce spring directions as auxiliary unknown variables. The system is globally linear in the node positions, and the non-linear terms involving the directions are strictly local. Because the global linear system does not depend on run-time state, the matrix can be pre-factored, allowing for very fast iterations. Our method converges to the same final result as would be obtained by solving the standard form of implicit Euler using Newtons method. Although the asymptotic convergence of Newtons method is faster than ours, the initial ratio of work to error reduction with our method is much faster than Newtons. For real-time visual applications, where speed and stability are more important than precision, we obtain visually acceptable results at a total cost per timestep that is only a fraction of that required for a single Newton iteration. When higher accuracy is required, our algorithm can be used to compute a good starting point for subsequent Newtons iteration.

Physics-inspired upsampling for cloth simulation in games

We propose a method for learning linear upsampling operators for physically-based cloth simulation, allowing us to enrich coarse meshes with mid-scale details in minimal time and memory budgets, as required in computer games. In contrast to classical subdivision schemes, our operators adapt to a specific context (e.g. a flag flapping in the wind or a skirt worn by a character), which allows them to achieve higher detail. Our method starts by pre-computing a pair of coarse and fine training simulations aligned with tracking constraints using harmonic test functions. Next, we train the upsampling operators with a new regularization method that enables us to learn mid-scale details without overfitting. We demonstrate generalizability to unseen conditions such as different wind velocities or novel character motions. Finally, we discuss how to re-introduce high frequency details not explainable by the coarse mesh alone using oscillatory modes.

A point-based method for animating elastoplastic solids

In this paper we describe a point-based approach for animating elastoplastic materials. Our primary contribution is a simple method for computing the deformation gradient for each particle in the simulation. The deformation gradient is computed for each particle by finding the affine transformation that best approximates the motion of neighboring particles over a single timestep. These transformations are then composed to compute the total deformation gradient that describes the deformation around a particle over the course of the simulation. Given the deformation gradient we can apply arbitrary constitutive models and compute the resulting elastic forces. Our method has two primary advantages: we do not store or compare to an initial rest configuration and we work directly with the deformation gradient. The first advantage avoids poor numerical conditioning and the second naturally leads to a multiplicative model of deformation appropriate for finite deformations. We demonstrate our approach on a number of examples that exhibit a wide range of material behaviors.

A point-based method for animating incompressible flow

In this paper, we present a point-based method for animating incompressible flow. The advection term is handled by moving the sample points through the flow in a Lagrangian fashion. However, unlike most previous approaches, the pressure term is handled by performing a projection onto a divergence-free field. To perform the pressure projection, we compute a Voronoi diagram with the sample points as input. Borrowing from Finite Volume Methods, we then invoke the divergence theorem and ensure that each Voronoi cell is divergence free. To handle complex boundary conditions, Voronoi cells are clipped against obstacle boundaries and free surfaces. The method is stable, flexible and combines many of the desirable features of point-based and grid-based methods. We demonstrate our approach on several examples of splashing and streaming liquid and swirling smoke.

A level-set method for skinning animated particle data

In this paper, we present a straightforward, easy to implement method for particle skinning-generating surfaces from animated particle data. We cast the problem in terms of constrained optimization and solve the optimization using a level-set approach. The optimization seeks to minimize the thin-plate energy of the surface, while staying between surfaces defined by the union of spheres centered at the particles. Our approach skins each frame independently while preserving the temporal coherence of the underlying particle animation. Thus, it is well-suited for environments where particle skinning is treated as a post-process, with each frame generated in parallel. Moreover, our approach is integrated with the OpenVDB library and the underlying partial differential equation is amenable to implicit time integration. We demonstrate our method on data generated by a variety of fluid simulation techniques and simple particle systems.

Physics-based animation of large-scale splashing liquids

Fluid simulation has been one of the greatest successes of physics-based animation, generating hundreds of research papers and a great many special effects over the last fifteen years. However, the animation of large-scale, splashing liquids remains challenging. In this paper, we show that a novel combination of unilateral incompressibility, mass-full FLIP, and blurred boundaries is extremely well-suited to the animation of large-scale, violent, splashing liquids.