Dear all, I am working on the subject of the simulation of Richtmyer-Meshkov instablity. The controllong equations are the compressible Euler/NS equations. At first, between the interfae there are two different fluid, which will be accelerated by a shock and then result in the so-called R-M instablity. Now, I am wondering what model I should use in my simulation. I am confused by the concept, such as multicomponent, multiphase and multifluid. So I hope someone could give me explaination about such fields. Thanks anyway!

Hi Tian,

I am not familiar with the Richtmyer-Meshkov instability but I may clarify the terminology multicomponent/multiphase.

- multicomponent: indicate the fluid (or liquid or solid) of interest is composed of several component. For example, air is a multicomponent gas composed essentially of O2 and N2.

- multiphase: a phase correspond to the state of the material you are investigated: gas, liquid, solid. A flow is said multiphased if it is composed of a mixing between two (or more) different phase. For example, the injection of fuel in an internal combustion engine is multiphase.

I will assume that multifluid notify if several fluid are interacting, but I am not familiar with the terminology.

Hoppe this will help. Julien

Hi, Tian, What jdc said is really right but for the Richtmyer-Meshkov instablity in your case, you don't need to consider the multicomponent, multiphase or any confusive concepts. No model is needed, as long as both gases are ideal. You just treat them as a single gas! What you need to do is specifying the initial field properly and then solving the Euler system.

- Paul

Thanks a lot to jdc and Paul. As for my problem,the R-M instablity,I have simulated the case in which both gas are ideal and treated as a single gas. Now,I want to investigate the effect of different state equation to R-M instablity,such as the interface of Air and He, Air and SF6, with different ratio of specific heats. So I think that I have to consider such problem,such as multicomponent, multiphase or multifluid. In literatures, some multicomponent models, even multiphase models were used to simulate the R-M instablity. In my simulation, the level set method was used to tract the moving interface. Moreover,I want to study the mixture of two fluids, so I want to know whether Euler equations are enough for the mixture or not. Is it that only the full N-S equations can describe the fluid mixture, even to turbulence?

best wishes!

Tian Baolin

Hi, Tian, at the very beginning of your research, you can still treat any combinations - Air/He, Air/SF6 or Ar/He - as a single gas, if the ratios of specific heat are supposed to be the same. Otherwise you have to solve for the equtions for the components, but then only multicomponent model is involved. Usually one would like to avoid such situation since then many things have to considered like the diffusive rate between two components. Good luck!

-Paul

Hi,Paul,thanks a lot. Maybe I will have to use the multicomponent model in my work, for my PhD supervisor asked me to investigate the R-M instablity with different ratios of specific heat. And I have finished the simulation as a single gas. I plan to combine the VOF and level set method to such multicomponent problem, but now I have not got the detailed solution. I don't want to adopt the Ghost method for I think it does not include the diffusion.

Another question, the diffusion accross the interface is of importance or not? What is true physical mechanics accross interface. In my opinion,both the level set method and the front tracking method fail to describe the diffusion, right?

Tian

It sounds like you are interested in questions that can be answered most directly by eliminating the front tracking and level sets, and going to a multicomponent treatment with the NS equations. The physics of the interface is that there is no true interface. As soon as you create a sharp interface, diffusion begins to smear it out. Whether diffusion is important depends on your length and time scales, transport coefficients and other physical parameters, and what questions you are trying to answer. Be sure to use enough resolution that numerical diffusion isn't a problem.