| The atomization of liquid fuel is a kind of violent dynamic process from continuous phase to discrete phase, involving complex fluid phenomena such as multiphase, turbulence, fluid interfaces etc. To understand atomization physics and mechanism in engineering application and academic regime, it is of great significance to develop an integrated numerical simulation method which is capable of simulating all sub-process in atomization consecutively. With such a motivation, an integrated multiphase numerical simulation frame, in which a mesh-less MPS (Moving particle semi-implicit method) is used to simulate the deforming of liquid fluid phase from continuous to discrete state, while the traditional FVM (finite volume method) based on Euler mesh is utilized to simulate the corresponding gas fluid phase movement with coupled two-way interaction, is newly proposed. However, the built-in instability of MPS algorithm has seriously limited its further development. In this dissertation, the stabilized3D MPS model and algorithm were investigated and programmed into an integrated multiphase simulation code, then the whole sub-processes of liquid injecting, break-up, atomization were simulated. The details are as following:(1) The error analysis of key arithmetic operators such as gradient operator for pressure gradient evaluation and Laplace operator for Poisson Equation in MPS is conducted theoretically and numerically based on one-dimensional analysis. It is founded that for both linear and non-linear functions, the main source of error comes from boundary particles. The error can be controlled by adjusting particle spacing distance or introducing dummy particles, eliminating its effects on instability.(2) A comprehensive comparison of stabilizing methods proposed by different researchers has been conducted. It is founded that most algorithms were validated based on2D and fine particle spacing conditions. For3D Case, it is yet dissatisfactory to obtain a stabilized computation by simply extending them to3D version, especially for coarse particle spacing conditions.(3) For further stability enhancement of MPS, a kind of artificial viscosity model is proposed, which is based on SPH (Smooth Particle Hydrodynamics) with extending and modification under the concept of MPS. It is validated that a combination usage of the artificial viscosity model as well those algorithms proposed previously plays an effective role in stability enhancement of MPS, while accuracy are guaranteed with a switching variable.(4) The stabilized3D MPS method is then applied in a series of integrated simulation of liquid atomization in quiescent and supersonic flow, the results are compared with those of existing methods such as VOF (Volume of fluid)å'ŒDPM(Discrete Phase Model). It is founded that the consecutive dynamic sub-processes of liquid injecting, deforming, break-up as well as atomization in quiescent and supersonic flow are computed successfully by proposed method, demonstrating its effectiveness and reliability, which is unique as compared with other existing methods.The innovation points of this dissertation mainly lie in:(1) In context of3D simulation, an artificial viscosity model based on SPH is proposed. Then with a combination usage of the artificial viscosity model as well those algorithms proposed previously, a relatively stabilized3D MPS model and method is obtained.(2) An integrated multiphase simulation method is coming into being by coupling the stabilized3D MPS with the traditional FVM based on Euler mesh, which is capable of simulating all sub-process in atomization integratively. There is no similar model and method reported in published literatures.(3) With newly developed methods and models, the liquid injection in supersonic air flow conditions are computed and the complex dynamic sub-processes from continuous to discrete state are captured successfully. There are also no similar results reported in published literatures. |
PDF Full Text Request