Implicit Density Projection for Volume Conserving Liquids.

Area of Science:

• Computer Graphics
• Computational Physics
• Fluid Dynamics

Background:

• Standard fluid simulation methods like FLIP and APIC struggle with volume conservation due to numerical errors.
• These errors accumulate over time, leading to unrealistic volume changes in simulations.

Purpose of the Study:

• To develop a novel method for enforcing volume conservation in hybrid Eulerian/Lagrangian fluid simulations.
• To address limitations of existing pressure solvers that cannot correct for accumulated volume errors.

Main Methods:

• Proposed an implicit density projection approach to enforce constant fluid density.
• Derived a pressure Poisson equation incorporating density deviations based on the mass conservation law.
• Integrated density tracking via simulation particles into existing solvers.

Main Results:

• Achieved robust volume conservation even in degenerate configurations and with boundary handling.
• Demonstrated significant improvements in incompressibility, visual realism, and computational performance compared to standard methods.
• Enabled relaxed time step and solver accuracy requirements, leading to higher performance.

Conclusions:

• The novel density projection method effectively maintains fluid volume and enhances simulation quality.
• The approach offers practical advantages, including better particle distribution and reduced need for resampling.
• This method represents a significant advancement for realistic and efficient fluid simulations.