Learning Forum

[shadowfax7]

I'm running steady-state simulations with injections of particles in a fluid domain (one-way coupling). As I increase particle size, less and less particles are able to exit the domain and are rendered 'incomplete'. I was wondering how I can make more particles exit the domain? With respect to tracking parameters, I've tried increasing the number of steps from 1 million to 2 million, but that doesn't seem to have any effect. n

[Surya Deb]

Hello, nDo you have gravity turned on or expect any kind of settling behavior with increased particle size?nAlso is there a noticeable change in the response time or Stokes number of the particles with a change in diameter?nDo you have Turbulent Dispersion/Random Walk turned on? Check if the increase in size leads to increased interaction with vortices or generation of additional turbulence.nWhere do the majority of particle tracks terminate? Maybe you can check the flow features in and around that region to identify the cause.nAlso check if the volume fraction of particles in cells get affected due to increased diameter. What is the max particle Volume fraction in the domain?nRegards,nSuryan

[YasserSelima]

I remember I read in earlier version of fluent (probably 12) manual that to simulate discrete phase and keep the mass balance, the simulation should be transient. For some mysterious reason fluent does not keep the mass balance of the discrete phase in some domains ... do not remember the details exactly and I am not sure if this fixed in later versions.

[YasserSelima]

Thanks to google!! does this apply to your case?.The steady-particle Lagrangian discrete phase model described in this chapter is suited for flows in which particle streams are injected into a continuous phase flow with a welldefined entrance and exit condition. The Lagrangian model does not effectively model flows in which particles are suspended indefinitely in the continuum, as occurs in solid suspensions within closed systems such as stirred tanks, mixing vessels, or fluidized beds. The unsteady-particle discrete phase model, however, is capable of modeling continuous suspensions of particles. See Chapters 22 and 24 for information about when you might want to use one of the general multiphase models instead of the discrete phase models.n

[Rob]

Correct for contained systems, however I think in this case the issue is the particles are settling and then don't want to move again. nif you plot the particle trajectories where do they go? The larger the particle the heavier it is, so the more likely it'll separate from the flow (if density is dissimilar: read up on Stokes Law). n

[shadowfax7]

Thanks all for your input! n- yes, gravity is enabled (see picture); no turbulence effects are modelled; for these type of particles the Stokes number is still below 1 according to my estimations; I haven't checked Volume Fraction because I don't exactly know how to do that, but I will check it - sounds interesting n(& all) - I've just looked at one particular particle track here, which suddenly seems to come to a halt and will likely be one of the non-exit particles (see arrow under 'Particle Tracks'). Looking at the fluid pathlines, and the contours of velocity magnitude in that region (see enlarged box), I think there is an increased stagnation zone near the bottom of the vessel due to the bend (?). In this set-up, gravity is enabled in the +Y direction (so downwards in the picture). It seems like the particle dislodges from the fluid stream (it does have a larger density: 1600 kg/m³ over 1060 kg/m³) and gets stuck in this region. nn

[Rob]

That looks like an artery. With that level of density difference and low speed regions with some swirl I'd expect some particles to get stuck in the domain. n

[YasserSelima]

Monitor the void fraction in this region. You can make an expression that gives 1 in the cells within certain x-y coordinates and zero every where else exp_variable. And use this to find void fraction in the region where particles disappear. nSum(CellVolume*void fraction* exp_variable )/Sum( CellVolume * exp_variable) nLet's see where does Fluent hide the particles

[shadowfax7]

Coming back to this, I've plotted the final position of each non-exit particle in the screenshot below. The region highlighted is the region I highlighted in the previous screenshot. It does seem - to me, at least - that particles are depositing in zones of low velocity near the bottom of the vessels (not always easy to see in 2D). nn

[DrAmine]

Is that deposition expected? Are they stuck so incomplete? Laminar?n

[shadowfax7]

hard to say whether it's expected or not, definitely will be running this analysis in transient because I think some of the particles getting stuck now would get an extra 'push' by pulsatile inflow, so I want to double-check that. Yes, these are the final positions of the 'incomplete' particles - so they seem stuck. Flow is laminar, indeed. n

[Rob]

They'll drop due to gravity and then get caught in the low velocity region near the walls. n