A NUMERICALLY CONSISTENT MULTIPHASE POISEUILLE FLOW COMPUTATION BY A NEW PARTICLE METHOD

2015 ◽  
Vol 76 (8) ◽  
Author(s):  
K. C. Ng ◽  
Y. H. Hwang ◽  
T. W. H. Sheu ◽  
M. Z. Yusoff

Recently, there is a rising interest in simulating fluid flow by using particle methods, which are mesh-free. However, the viscous stresses (or diffusion term) appeared in fluid flow governing equations are commonly expressed as the second-order derivatives of flow velocities, which are usually discretized by an inconsistent numerical approach in a particle-based method. In this work, a consistent method in discretizing the diffusion term is implemented in our particle-based fluid flow solver (namely the Moving Particle Pressure Mesh (MPPM) method). The new solver is then used to solve a multiphase Poiseuille flow problem. The error is decreasing while the grid is refined, showing the consistency of our current numerical implementation.

Author(s):  
Kailun Guo ◽  
Ronghua Chen ◽  
Suizheng Qiu ◽  
Wenxi Tian ◽  
Guanghui Su ◽  
...  

Multiphase flow widely exists in the nature and engineering. The two-phase flow is the highlight of the studies about the flow in the vessel and steam explosion in nuclear severe accidents. The Moving Particle Semi-implicit (MPS) method is a fully-Lagrangian particle method without grid mesh which focuses on tracking the single particle and concerns with its movement. It has advantages in tracking complex multiphase flows compared with gird methods, and thus shows great potential in predicting multiphase flows. The objective of this thesis is to develop a general multiphase particle method based on the original MPS method and thus this work is of great significance for improving the numerical method for simulating the instability in reactor severe accident and two-phase flows in vessel. This research is intended to provide a study of the instability based on the MPS method. Latest achievements of mesh-free particle methods in instability are researched and a new multiphase MPS method, which is based on the original one, for simulating instability has been developed and validated. Based on referring to other researchers’ papers, the Pressure Poisson Equation (PPE), the viscosity term, the free surface particle determination part and the surface tension model are optimized or added. The numerical simulation on stratification behavior of two immiscible flows is carried out and results are analyzed after data processing. It is proved that the improved MPS method is more accurate than the original method in analysis of multiphase flows. In this paper, the main purposes are simulating and discussing Rayleigh-Taylor (R-T) instability and Kelvin-Helmholtz (K-H) instability. R-T and K-H instability play an important role in the mixing process of many layered flows. R-T instability occurs when a lower density fluid is supported by another density higher fluid or higher density fluid is accelerated by lower density fluid, and the resulting small perturbation increases and eventually forms turbulence. K-H instability is a small disturbance for two different densities, such as waves, at the interface of the two-phase fluid after giving a fixed acceleration in the fluid. Turbulence generated by R-T instability and K-H instability has an important effect in applications such as astrophysics, geophysics, and nuclear science.


Author(s):  
Rida S. N. Mahmudah ◽  
Masahiro Kumabe ◽  
Takahito Suzuki ◽  
LianCheng Guo ◽  
Koji Morita ◽  
...  

Understanding the freezing behavior of molten metal in flow channels is of importance for severe accident analysis of liquid metal reactors. In order to simulate its fundamental behavior, a 3D fluid dynamics code was developed using Finite Volume Particle (FVP) method, which is one of the moving particle methods. This method, which is fully Lagrangian particle method, assumes that each moving particle occupies certain volume. The governing equations that determine the phase change process are solved by discretizing its gradient and Laplacian terms with the moving particles. The motions of each particle and heat transfer between particles are calculated through interaction with its neighboring particles. A series of experiments for fundamental freezing behavior of molten metal during penetration on to a metal structure was also performed to provide data for the validation of the developed code. The comparison between simulation and experimental results indicates that the present 3D code using the FVP method can successfully reproduce the observed freezing process such as molten metal temperature profile, frozen molten metal shape and its penetration length on the metal structure.


Geofluids ◽  
2020 ◽  
Vol 2020 ◽  
pp. 1-14
Author(s):  
D. Roubinet ◽  
S. Demirel ◽  
E. B. Voytek ◽  
X. Wang ◽  
J. Irving

Modeling fluid flow in three-dimensional fracture networks is required in a wide variety of applications related to fractured rocks. Numerical approaches developed for this purpose rely on either simplified representations of the physics of the considered problem using mesh-free methods at the fracture scale or complex meshing of the studied systems resulting in considerable computational costs. Here, we derive an alternative approach that does not rely on a full meshing of the fracture network yet maintains an accurate representation of the modeled physical processes. This is done by considering simplified fracture networks in which the fractures are represented as rectangles that are divided into rectangular subfractures such that the fracture intersections are defined on the borders of these subfractures. Two-dimensional analytical solutions for the Darcy-scale flow problem are utilized at the subfracture scale and coupled at the fracture-network scale through discretization nodes located on the subfracture borders. We investigate the impact of parameters related to the location and number of the discretization nodes on the results obtained, and we compare our results with those calculated using reference solutions, which are an analytical solution for simple configurations and a standard finite-element modeling approach for complex configurations. This work represents a first step towards the development of 3D hybrid analytical and numerical approaches where the impact of the surrounding matrix will be eventually considered.


Author(s):  
LianCheng Guo ◽  
Shuai Zhang ◽  
Koji Morita ◽  
Kenji Fukuda

Sloshing dynamics of a molten core is one of the fundamental behaviors in core disruptive accidents of a liquid-metal cooled reactor. In addition, solid particle-liquid mixture comprising molten fuel, molten structure, refrozen fuel, solid fuel pellets, etc. could lead to damping of its flowing process in a disrupted core. The objective of the present study is to investigate the applicability of the finite volume particle method (FVP), which is one of the moving particle methods, to 3D motion of liquid sloshing processes measured in a series of experiments. In the first part of this study, a typical sloshing experiment of single liquid phase is simulated to verify the present 3D FVP method for sloshing characteristics that include free surface behaviors. Second, simulations of sloshing problems with solid particles are performed to validate the applicability of the FVP method to the 3D motion of solid particle-liquid mixture flows. Some good agreements between the simulation and its corresponding experiment demonstrate applicability of the present FVP method to 3D fluid dynamics of liquid sloshing flow with solid particles.


2018 ◽  
Vol 21 (2) ◽  
pp. 279-294 ◽  
Author(s):  
Ehsan Jafari Nodoushan ◽  
Ahmad Shakibaeinia

Abstract Sediment scouring is a common example of highly dynamic sediment transport. Considering its complexities, the accurate prediction of such a highly dynamic multiphase granular flow system is a challenge for the traditional numerical techniques that rely on a mesh system. The mesh-free particle methods are a newer generation of numerical techniques with an inherent ability to deal with the deformations and fragmentations of a multiphase continuum. This study aims at developing and evaluating a multiphase mesh-free particle model based on the weakly compressible moving particle semi-implicit (WC-MPS) formulation for simulation of sediment scouring. The sediment material is considered as a non-Newtonian viscoplastic fluid, whose behavior is predicted using a regularized μ(I) rheological model in combination with pressure-dependent yield criteria. The model is first validated for a benchmark problem of viscoplastic Poiseuille flow. It is then applied and evaluated for the study of two classical sediment scouring cases. The results show that the high-velocity flow currents and the circulations can create a low-viscosity region on the surface of the sediment continuum. Comparing the numerical results with the experimental measurements shows a good accuracy in prediction of the sediment profile, especially the shape and dimensions of the scour hole.


2016 ◽  
Vol 715 ◽  
pp. 198-202
Author(s):  
Ryota Shimono ◽  
Keiko Watanabe

The phenomena that occur during high-speed penetration of a projectile into sand particles are interesting subjects in engineering. The macro-scale research themes are the behavior of the ejected sand particles and the progress of the high-speed projectile, while the micro-scale research themes are the deformation and fragmentation of a single sand particle. Studies of these unique phenomena were conducted using both experiments and numerical simulation. Although accurate simulation of the behavior of sand particles during high-speed penetration is difficult because sand particles have characteristics of both fluids and solids, the reproducibility of the actual phenomena has improved in recent years with the development of particle methods. In our research, we conducted simulations of the phenomena using Smoothed Particle Hydrodynamics (SPH), which is a mesh-free, particle-based method. The results showed the possibility of accurate reproduction during high-speed projectile penetration into sand particles at the macro-scale.


Author(s):  
Eiji Ishii ◽  
Taisuke Sugii

The particle method is a useful approach to simulate fluid flows within micro/nano spaces such as micro-electromechanical systems, nano-in-print processes, and head-disk interfaces of hard disk drives. Particle methods are based on continuum dynamics, and some studies have recently extended the scope of these methods to approaches within micro/nano spaces. Surface tension is a dominant force in the fluid flow within micro/nano spaces. However, surface-tension models used in the particle methods need to be improved to achieve more stable and accurate simulation. In the present study, we developed a new surface tension model for the particle method using inter-particle force to improve the stability and accuracy of simulation; the inter-particle force was given by the derivation of potential energy in space. The developed surface tension model was verified using simple benchmark tests: pressure in a round droplet and oscillation period of a square liquid-droplet. The predicted pressure in a round droplet agreed well with that given by the Young-Laplace equation, and the predicted oscillation period of a square droplet agreed well with that given by Lamb’s theory. The wall-adhesion was also verified at various contact angles; heights of droplets on the wall agreed well with those given theoretically. We found that our new surface tension model was useful for simulating fluid flow within micro/nano spaces for particle method.


2016 ◽  
Vol 20 (3) ◽  
pp. 660-688 ◽  
Author(s):  
Andrea Colagrossi ◽  
Emanuele Rossi ◽  
Salvatore Marrone ◽  
David Le Touzé

AbstractIn this work two particle methods are studied in the context of viscous flows. The first one is a Vortex Particle Method, called Diffused Vortex Hydrodynamics (DVH), recently developed to simulate complex viscous flows at medium and high Reynolds regimes. This method presents some similarities with the SPH model and its Lagrangian meshless nature, even if it is based on a different numerical approach. Advantages and drawbacks of the two methods have been previously studied in Colagrossi et al. [1] from a theoretical point of view and in Rossi et al. [2], where these particle methods have been tested on selected benchmarks. Further investigations are presented in this article highlighting analogies and differences between the two particle models.


2011 ◽  
Vol 49 (5) ◽  
pp. 649-656 ◽  
Author(s):  
Aidin Jabbari Sahebari ◽  
Yee-Chung Jin ◽  
Ahmad Shakibaeinia

Sign in / Sign up

Export Citation Format

Share Document