Abstract
A three-dimensional full-scale centrifugal pump is simulated using the moving particle semi-implicit method (MPS). The genetic smooth wall boundary (GSW) is used and extended to 3D to deal with the complicated wall shape and thin-walled structure such as blades in turbo-machines. The non-surface detection technique (NSD) based on conceptual particles is combined with this wall model to avoid the loss of particle-number-density near wall boundaries. A local wall particle refinement method is developed. Fine particles and coarse particles are applied to curved surface and flat surface respectively so as to reduce the computational cost while maintain the high discretizing precision. The fully developed velocity inflow and pressure outflow boundary conditions are proposed. Two hydrostatic cases with complicated geometry are tested to verify the proposed models. This paper constructs a framework for the simulation of incompressible fluid machines, in which 3D complex revolving bodies can be integrally discretized and interactions between the stator and rotor can be integrally solved within a single coordinate. The results are compared with those of the fine volume method and good agreement is achieved. This paper provides a particle-based solver for incompressible fluid machinery and has the potential to study its inner flow with multiphase or phase change.
| Original language | English |
|---|---|
| State | Published - 2021 |
| Event | 7th International Conference on Particle-Based Methods: Fundamentals and Applications, PARTICLES 2021 - Hamburg, Germany Duration: 4 Oct 2001 → 6 Oct 2001 |
Conference
| Conference | 7th International Conference on Particle-Based Methods: Fundamentals and Applications, PARTICLES 2021 |
|---|---|
| Country/Territory | Germany |
| City | Hamburg |
| Period | 4/10/01 → 6/10/01 |
Keywords
- Centrifugal Pumps
- MPS method
- Open Boundary conditions
- Wall Boundary Conditions
Fingerprint
Dive into the research topics of 'SIMULATION OF A CENTRIFUGAL PUMP BASED ON THE MESHLESS SOLVER'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver