A Variational Staggered Particle Framework for Incompressible Free-Surface Flows
Smoothed particle hydrodynamics (SPH) has been extensively studied in computer graphics to animate fluids with versatile effects. However, SPH still suffers from two numerical difficulties: the particle deficiency problem, which will deteriorate the simulation accuracy, and the particle clumping problem, which usually leads to poor stability of particle simulations. We propose to solve these two problems by developing an approximate projection method for incompressible free-surface flows under a variational staggered particle framework. After particle discretization, we first categorize all fluid particles into four subsets. Then according to the classification, we propose to solve the particle deficiency problem by analytically imposing free surface boundary conditions on both the Laplacian operator and the source term. To address the particle clumping problem, we propose to extend the Taylor-series consistent pressure gradient model with kernel function correction and semi-analytical boundary conditions. Compared to previous approximate projection method [1], our incompressibility solver is stable under both compressive and tensile stress states, no pressure clumping or iterative density correction (e.g., a density constrained pressure approach) is necessary to stabilize the solver anymore. Motivated by the Helmholtz free energy functional, we additionally introduce an iterative particle shifting algorithm to improve the accuracy. It significantly reduces particle splashes near the free surface. Therefore, high-fidelity simulations of the formation and fragmentation of liquid jets and sheets are obtained for both the two-jets and milk-crown examples.