The traditional SPH fluid simulation method usually uses fixed particle smooth length for interpolation, which may result in large interpolation errors in some cases. In order to improve the simulation accuracy, this paper established the relationship between the smoothing length of particles and the harmonic mean of neighboring particle densities, and adaptively adjusted the smoothing length of particles. Furthermore, the paper designed the corresponding particle neighbor search scheme and kernel function to solve the problem of force asymmetry. Experiments show that this method effectively reduces the variance of the number of particle neighbors, which solves the problem of large interpolation error caused by fixed support domain of traditional SPH particles, and makes the simulation results closer to physical facts. At the same time, due to improved calculation accuracy, the simulation stability is enhanced, and the physical simulation can be performed with a larger time step to effectively improve the simulation efficiency. Finally, compared with state-of-the-art methods, this method has advantages both in visual quality and simulation speed.

[2020-09-05] Printed Article /doc/2019.06.0172/v2