[czhao86]

When I turn on the Brownian force for DPM, the particles are escaping from the inlet, which does not make sense as there is a velocity-inlet going towards the outlet. Anyone have some idea?

[diamondx]

what injection model are you using ?

[czhao86]

I am using surface injection at inlet now.

I am trying the file injection to check whether that is the problem or not.

[flotus1]

It does make sense that the particles escape from the inlet.

With Brownian motion, the particles behave like molecules which means that they have a velocity distribution function according to a Maxwell-Boltzmann equilibrium distribution function.

The smaller the particles, the higher the velocity of the Brownian motion. This velocity can become higher than the velocity of the flow, leading to particles travelling against the flow direction and leaving the domain at the inlet.

[czhao86]

Yes, I think maybe that is the case, but when I searched cfd-online, it seems like a group of people don't like the Brownian motion in Fluent, and said the equation is not right in some case.

I am trying to know whether that is the case, then I need to write a User defined brownian force, especially after I read the paper: http://dx.doi.org/10.1080/02786820701203223

When I put particles in a volume and turn on the Brownian motion in Fluent, it looks fine. So maybe surface injection is not good with Brownian motion on.

Thank you very much.

[flotus1]

I dont know what causes the error in the Brownian motion model observed in the cited paper in fluent.
The formulas in the manual seem to be correct, at least if an implicit or explicit Euler method is uses for discretization.