| Thread Previous • Date Previous • Date Next • Thread Next |
Thank you for your answer. So, if I understand well, the RPM is quite similar to the CohesiveFrictionalContactLaw? I don't grab all the idea. Could you please be a bit more illustrative?A few words about my own work ;-) Check out ConcretePM as well. It is more complicated than RPM I think, but the basic idea is the same; there is a high-level overview in comment at the beginning of ConcretePM.hpp, which you might find useful to get you started with your own constitutive law; also read http://yade.wikia.com/wiki/ConstitutiveLawHowto.
The difference of ConcretePM/RockPM (to, say, CohesiveFrictionalContactLaw) is that geometry computation is offloaded to the interactionPhysics class (Dem3DofGeom::displacementN() etc), so you have only the constitutive law itself at one place, in form that is quite similar to its mathematical formulation. That makes writing new constitutive laws easier.
The idea is to create some particles "connected" (cohesive) with each other, if they are close to each other on the first stage of simulation.In ConcretePM, cohesive contacts are only create in the first timestep, later it is only non-cohesive (which use the same constitutive law, but with damage initialized to 1 instead of 0) that will occur.
ESyS-Particle, on the other hand, does it a different way, they load sphere packing as well as initial bonds externally and only deal with creating non-cohesive interactions (contacts) as spheres bump into each other.
Cheers, Vaclav
| Thread Previous • Date Previous • Date Next • Thread Next |
