[Aarthy Meena]

Hi all,

I am trying to model natural convection heat transfer of nanofluids inside a square cavity using DPM method. I have following queries regarding that

1) How to convert volume fraction of nanoparticles to number of parcels in case of group injections? For eg: if I want to use 2% volume fraction of nanoparticles of 30 nm diameter, how to create injections for this case?
2) As it is natural convection, initially I need to define a uniform distribution of nanoparticles through out the flow domain.

I strongly believe that, experts here can clarify my doubts. Thank You.

[mprinkey]

(1) To provide fix volume fraction at an inlet requires that both phases enter at the same inlet in relative volume proportions. That is just simple algebra. You may need to think if there is a slip velocity between the two phases, but the starting point should be to assume the particle and fluid share the same velocity.

(2) You will need to a patch in DPM particles in the flow. But, you will not know the fluid or particle velcocities before hand, so you will potentially have settling issues as the convection develops. You might be just as well off to start with no initial particles and just run the simulation until it reaches the target volume fraction. Note that if you have an outlet (with or without a screen) and a particle inlet, you cannot enforce a full field average volume fraction. Even if you have no outlet, if you have a particle injection, the particle population will grow with time.

If the cavity is truly closed and you are seek steady state results, I'd question your rationale about using particle injections at all. Just do initial particle distributions at your target volume fraction and iterate until steady state...or at least until you can time average a transient window.

[Aarthy Meena]

Originally Posted by mprinkey

(1) To provide fix volume fraction at an inlet requires that both phases enter at the same inlet in relative volume proportions. That is just simple algebra. You may need to think if there is a slip velocity between the two phases, but the starting point should be to assume the particle and fluid share the same velocity.

(2) You will need to a patch in DPM particles in the flow. But, you will not know the fluid or particle velcocities before hand, so you will potentially have settling issues as the convection develops. You might be just as well off to start with no initial particles and just run the simulation until it reaches the target volume fraction. Note that if you have an outlet (with or without a screen) and a particle inlet, you cannot enforce a full field average volume fraction. Even if you have no outlet, if you have a particle injection, the particle population will grow with time.

If the cavity is truly closed and you are seek steady state results, I'd question your rationale about using particle injections at all. Just do initial particle distributions at your target volume fraction and iterate until steady state...or at least until you can time average a transient window.

Dear Sir,

Thank you very much for your elaborate reply.

I completely understood the point 2 of your post. Thanks for giving deep insights.

Sir, Can you please elaborate more on point 1. For eg: Let us take a scenario of flow around a circular cylinder in cylindrical coordinates with forced flow in laminar steady regime. I need a 2% volume fraction nanofluid to flow.

1) How will I decide the type of injection? single or surface or group.
2) How will I control the number of parcels, so that my volume fraction requirement is met out?

Sorry, I understand that you said it is a simple algebra. But, it will be helpful if you can give some more hints or a reference material which will be helpful in understanding these concepts.

Thank you once again.

[mprinkey]

Volume fraction is defined as the fraction of the total volume occupied by a given phase. You can compute the volume of a sphere for a given particle radius, right? And the number of particles times that particle volume divided by the total volume. That is all it takes. The details will depend on your particular solver input.

[Aarthy Meena]

Originally Posted by mprinkey

Volume fraction is defined as the fraction of the total volume occupied by a given phase. You can compute the volume of a sphere for a given particle radius, right? And the number of particles times that particle volume divided by the total volume. That is all it takes. The details will depend on your particular solver input.

Dear Sir,

Thank you so much for your kind reply. I will update you with the progress. Thank You :-)