Dear all,
firstly as a new user I like to say Hello!
Although I understand the ideas of DEM pretty well, my programming skills with respect to C++ are practically in-existent. From what I understood by reading the manual and 1 or 2 entries in the Forum, my problem may need some programming, but please correct me, if I am wrong.
My project is the modeling of mixing of cuboid shaped flat particles with hollow spheres (dimensions are in the 10's mm range) in a rotary drum of defined geometry. The spheres are hollow to account for the density difference between the cuboid shape particles and the spherical particles. The materials have density of about 1500 to 2500 for the cuboids and about 8000 for the spheres.
As I understand from bidisperse spherically shaped particle mixtures, they can segregate if a certain size or density ratio is present. Now the data from the latter can to my understanding not be easily transferred to my set-up due to the differences in shape. Also, I cannot do a purely experimental investigation due to costs. My supervisor will not spend money on hollow spheres of different diameters and wall thicknesses without any theoretical results that suggest a parameter range.
My approach is therefore:
1) Show segregation of a bidisperse mixture of equally sized spheres of different densities. (The hollow spheres are actually hollow because I need to adapt the sphere material density to the density of the cuboid shape particles.) This should be fairly simple I think. There are examples in the liggghts installation that can be adapted to do that.
2) Show that segregation will not occur for equally sized spheres, while one is hollow. Now here is my first "problem" with liggghts:
I cannot define a hollow sphere. In an older forum entry (Sorry, I cannot find it anymore.) it was suggested to adapt the density of the sphere to account for its hollowness. Now this will of course be an easy work-around but in my eyes it causes an error in the conservation of angular momentum equation for the particle, since the moment of inertia of a full sphere of reduced density and of a hollow sphere of real density are not equal. I could also adapt the hollow sphere density to equalize the inertial moments but then the gravitational force is incorrect. You probably see the dilemma.
Unfortunately I have no clue on how to solve the problem, especially if programming skills are required. But maybe one of you could at least suggest, where I need to look and where to do changes. I guess one would need to adapt the spherical particles that are there and give them a second inner radius. At this point I might also ask, in what what the moment of inertia of the spheres is computed?
For the moment i assume this problem can be solved so the next step would be showing the segregation for different hollow spheres of different size compared to the full spheres (that represent the cuboid particles).
3) Step 3 is replacing the full spheres with cuboid shaped particles and do the parameter study (mixture with differently sized hollow spheres) again to find a proper size of the hollow spheres so that segregation in a mixture of the cuboids and hollow sphere is as low as possible.
Here my problem is of course that the particles are no spheres. I read in the forum that it is possible to build the cuboid shape by using the multisphere-approach,meaning I stick multiple spheres together and create the shape I want. Since I use ligggghts in the PUBLIC version I have to do that on my own, so defining a sphere of size and density of the cuboid material and stick them together to form it. But I actually don't understand from the manual how I do that and also if that is the only and simplest approach. I also wonder wether or not I could define a particle shape via an STL file as it would be done with the rotary drum geometry.
If the problem is solved, I of course can do a parameter study with differently sized spheres (variation of radius and wall thickness) and derive some reasonable size for the hollow spheres.
Maybe some of you can give me an idea on what to do and which way to take to achieve my goal.
Finally it is my job to thank you in advance for any help or advice you can provide!
Best regards,
André.