[afo]

Dear Foamers,

I'm simulating the charging of a hopper, whose shape you can see in the attached files, through icoUncoupledKinematicParcelFoam solver, and I have come along with a strange behaviour by some of the particles: while descending, a few of them, instead of impacting the bottom, start to ascend stuck to the the wall and spinning around the hopper itself. Can this be due to the cylindrical shape of the hopper, and as a result I shall adopt a cylindrical or polar coordinate system? I'm telling this since whether I run a simulation with the same parameters, on a slice of the hopper, that is without a cylindrical symmetry, this behaviour doesn't occur.
If anybody has a hint or has encountered the same problem of mine, I will be grateful to her or him who can help.
Cheers

Paolo

[mxylondon]

I have the same trouble now. Waiting for answer.

[afo]

Dear Mxylondon, I solved this problem by setting properly the material's parameters, cylindrical coordinates are not necessary. I don't know if you might have direct experience with granular materials, this behaviour depends mainly from the elastic forces between particles, their bulk density and diameter; in deed I tried to discharge some alumina spheres from a bottle into a funnel, and I noticed that at the bottom of the funnel this unexpected behaviour actually occured, though it was less prominent. I suppose that the collision between the alumina spheres is so elastic that some particles may acquire a substantial amount of energy, and because of the combination of the collisions, they enter in a spinning and ascending trajectory, as a result they need some time to dissipate that, since friction at the funnel's wall is close to be null. In my case study, alumina is not the material of interest, actually I studied coke for blast furnace, coke has quite inelastic behaviour, both with its own particles and with walls, wall friction is quite high, so I set the parameters until the simulation showed a behaviour similar to what I experienced directly in the field of operation. I don't know what material you have to test, so I'm not sure whether my solution is suitable, by the way, try to give an inelastic collision at wall, just to try, and see what happens.

[afo]

If you don't know, in order to set the inelastic collision at wall, open the kinematicCloudProperties file of your case, at it's bottom you should find a section called standardWallInteractionCoeffs, there you have as deafult:

standardWallInteractionCoeffs
{
type rebound;
}

type this to set the condition cited above

{
type rebound;

mu 0;
}

where mu is the coefficient of restitution, it varies from 0, inelastic, to 1, perfectly elastic. Try to tune it according to your actual condition, and see what happens.