Woosuck Hong

Woosuck Hong's web page

All pictures in this web are my own work, which were produced using C++ and OpenGL.

About me

Ph.D candidate in Department of Computer Science, Texas A&M University under the supervision of Dr. Donald H. House and Dr. John Keyser

Contact : wshong@cs.tamu.edu

Spring 2008

CPSC 206(Section 504-505) : Click here

Research Interests:

Fluid simulation, Physically based modeling, Simulation of natural phenomena, Data-driven approach (Character animation), and Computer animation

My DemoReel[divx](Please click here to watch my movie)

A. Eulerian fluid simulation /See Showreel Chapter 1

Eulerian fluid simulations generally solve the Navier-Stokes equation over a grid. We can achieve full 3D liquid simulation based on the Navier-Stokes equations for incompressible flow. In this example I used a fixed finite-difference grid(the Marker and Cell method) for computing partial spatial derivatives, an implicit integration scheme for advection and diffusion which makes significant improvements on the stability of this approach to fluid simulation using a semi-Lagrangian scheme to handle the convection term and A preconditioned conjugate gradient method was used to solve a Poisson equation for fast and more accurate computation of pressure to achieve divergence free flow for incompressibility. In order to handle the fluid-air interface I used particle level set methods which assigns additional marker particles outside the fluid volume to enhance the accuracy of the fluid interface. For velocity extrapolation, I compute an implicit representation of the fluid surface directly from the particles, and then use the Fast Sweeping Method, which is O(n) in time to construct a signed distance field and faster than the Fast Marching Method. Ray tracing technique was used for rendering this simulation.

B. Hybrid fluid simulation /See Showreel Chapter 2

For this example movie, I implemented Bridson’s work which exploit the hybrid fluid simulation approaches called Particle-in-Cell (PIC) and Fluid-Implicit-Particle (FLIP). One of the drawbacks of a particle based fluid simulation is the O(n2) time complexity of the calculation of particle interactions. To circumvent this, PIC and FLIP employ a grid for intermediate calculations. Both approaches are fundamentally particle-based, as particles carry the momentum of the fluid, and the grid is used mainly for the efficient computation of the spatial interactions required to compute diffusion and guarantee incompressibility. Using this auxiliary grid, incompressibility and boundary conditions can be enforced much more efficiently than in a pure Lagrangian scheme like MPS. In both PIC and FLIP, mass particles have their own velocity and position, which are integrated numerically using velocity updates obtained from the grid. Since the methods are Lagrangian, the advection step of the semi-Largrangian method can be avoided, greatly reducing numerical dissipation and loss of flow detail. Marching cube technique was used after the implicit function was computed. The grid resolution for surface generation was 150 × 150 × 150 and 623,575 particles were generated in the initial step.

C. Lagrangian fluid simulation /See Showreel Chapter 3

1. Smoothed particle hydrodynamics (SPH)

Smoothed particle hydrodynamics (SPH) developed by Lucy , was designed to simulate astrophysical phenomena. The fluid is composed of a set of particles with inter-particle forces such as pressure and viscosity which can be computed at the position of a particle by a smoothing kernel. In this example, I implemented Miller’s work which extended the SPH based technique by Desbrun in order to simulate highly deformable objects for interactive fluid simulation. Marching cube technique was used for the surface after the implicit function was computed. 2000 particles was generated for this example.

2. Moving-Particle Semi-Implicit Method (MPS)

For this example movie, I implemented Premoze’s work which introduced Koshizuka’s original work on the Moving-Particle Semi-Implicit Method (MPS) to simulate incompressible liquid and multifluid flow. Unlike SPH fluid simulation, which is inherently compressible, MPS fluid simulation is incompressible, and thus creates more realistic looking water motion.

D. My interest on fluid research.

We propose a particle-based technique for simulating incompressible fluid that integrates adaptive refinement of particle sampling. Each particle represents a mass of fluid in its local region. Particles can be split into several particles in highly deformable areas for a detailed representation. Similarly, in regions of smooth flow, nearby particles can be merged into a single particle. Depth below the surface and Reynolds number are exploited as our criteria
for determining whether splitting or merging should take place. For the fluid dynamics calculations, we use the hybrid FLIP method, which is computationally simple and efficient.
Each particle has a volume and an effective range, based on its volume. A kernel function is used for transferring and updating the particle’s physical properties such as mass and velocity. Our adaptive particle-based simulation is demonstrated in several scenarios, demonstrating its effectiveness in capturing fine detail of the flow, where needed, while efficiently sampling regions where less detail is required.