人工黏性增强三维移动粒子半隐算法稳定性
ENHANCEMENT OF STABILITY OF 3D MOVING PARTICLE SEMI-IMPLICIT METHOD BY ARTIFICIAL VISCOSITY
-
摘要: 通过综合比较移动粒子半隐方法(moving-particle semi-implicitmethod, MPS)各种稳定性计算方法, 考虑了多种可能导致计算不稳定的因素, 并首次提出加入人工黏性来抑制非物理压力振荡, 得到了较为稳定的三维MPS算法.采用各种稳定性方案对比模拟了三维立方流体旋转状态下的变形及三维静水压问题, 并进一步与商业软件的流体体积函数方法(volumeof fluid, VOF)模型计算结果对比, 验证了其正确性.结果表明:对三维情形, 单纯应用已有的稳定性算法仍难以满足MPS计算稳定性的要求, 而进一步辅以该文提出的人工黏性方法则可以在准确性及稳定性方面均取得较好的效果, 且计算结果显示, 通过该文提出的开关模式增加人工黏性并不影响压强泊松方程的求解精度, 因而不会对流场的求解产生负面作用.Abstract: By the virtue of Lagrangian gridless particle method, the moving particle semi-implicit (MPS) method has been applied in a wide range of engineering applications. However, the built-in instability in computational algorithm has seriously limited its further development and application. In this paper, aiming at developing 3D stabilized MPS method, a comprehensive comparison of stabilizing methods proposed by different researchers has been conducted, in which many elements leading to instability are considered. Then the equations proposed in 2D are extended to 3D and an artificial viscosity model based on smooth particle hydrodynamics (SPH) is proposed to reduce instability in MPS for the first time. With the different combinations of proposed methods and models, the deforming process of a swirling cubic fluid patch and hydrostatic pressure problem are simulated comparatively. The results are also compared with those from volume of fluid (VOF) model conducted by a commercial software. The accuracy and stability of proposed methods and models are validated. It is concluded that: for 3D case, only extending and applying previous stabilizing method is still dissatisfactory, while good effects in accuracy and stability can be obtained by incorporating the artificial viscosity model proposed in present paper. And it is also revealed that no negative effects on exactness of pressure solution will be produced, and thus on the flow field.