A boundary integral based particle initialization algorithm for Smooth Particle Hydrodynamics

Parikshit Boregowda,G R Liu

2024-08-15

Abstract:Algorithms for initializing particle distribution in SPH simulations are essential for enhancing simulation accuracy. However, no such algorithms exist for boundary integral SPH models, which can model complex geometries without requiring layers of virtual particles. This study introduces the Boundary Integral based Particle Initialization (BIPI) algorithm. It employs a particle shifting technique meticulously designed to redistribute particles to fit the geometry boundary. The BIPI algorithm directly utilizes the geometry's boundary information using the SPH boundary integral formulation. Special consideration is given to particles adjacent to the boundary to prevent artificial volume compression. Consequently, it can automatically generate a "uniform" particle distribution with reduced and stabilized concentration gradients for domains with complex geometrical shapes. Finally, several examples are presented to demonstrate the effectiveness of the proposed algorithm, including the application of the BIPI algorithm in flow problems.

Computational Engineering, Finance, and Science,Numerical Analysis

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: in Smooth Particle Hydrodynamics (SPH) simulations, how to provide an effective particle initialization algorithm for the boundary - integral SPH model to achieve a uniform particle distribution under complex geometries. Specifically, the paper aims to:

1. **Avoid the use of virtual particles**: Traditional methods usually need to add a virtual particle layer to handle boundary problems, which becomes very complicated under complex geometries. The paper proposes a method that does not rely on virtual particles.

2. **Improve simulation accuracy**: By introducing the Boundary Integral - based Particle Initialization (BIPI) algorithm, ensure a more uniform particle distribution, thereby reducing concentration - gradient fluctuations and improving the accuracy of SPH simulations.

3. **Optimize the particle initialization process**: Existing particle initialization methods are prone to cause artificial volume compression or particle rearrangement when dealing with the vicinity of boundaries, affecting the simulation results. The BIPI algorithm ensures a reasonable distribution of particles near the boundaries through a carefully - designed particle movement technique, avoiding these problems.

4. **Adapt to complex geometries**: The BIPI algorithm proposed in the paper can directly utilize geometric boundary information and is applicable to computational domains of various complex shapes, such as boundaries represented by line segments and triangles in two - dimensional and three - dimensional spaces.

In summary, the core objective of the paper is to develop a new particle initialization algorithm that can effectively handle boundary problems under complex geometries and does not require a virtual particle layer, thereby improving the overall accuracy and efficiency of SPH simulations.