Abstract
The Moving Particle Semi-implicit method (MPS) shows great potential to simulate violent incompressible free surface flows. However, Direct Numerical Simulations of such turbulent flows by particle methods would be extremely time-consuming. Taken ideas from the Sub-Grid Scale (SGS) model in grid-based Large Eddy Simulation (LES), static and dynamic Smagorinsky models are specially designed for Sub-Particle Scale (SPS) model in MPS. These SPS models are discretized by the combination of gradient and divergence models in MPS and the semi-implicit algorithm to implement MPS-LES is developed. The widely-studied mixing layer flows are simulated to quantitatively validate the developed turbulence particle model. Both the turbulent statistical profiles and vortex pairing processes are in good agreements with experimental observations after the SPS model is adopted. Moreover, it is found that the performance of the dynamic Smagorinsky model is better than that of the static model.
Original language | English |
---|---|
Pages (from-to) | 3135-3149 |
Number of pages | 15 |
Journal | Applied Mathematical Modelling |
Volume | 39 |
Issue number | 10-11 |
DOIs | |
Publication status | Published - 2015 Jan 1 |
Externally published | Yes |
Keywords
- Large Eddy Simulation
- Particle method
- Sub-Particle-Scale model
- Turbulent mixing layer flow
ASJC Scopus subject areas
- Modelling and Simulation
- Applied Mathematics