A simple question on interactionDetectionFactor

Hello,

We mentioned intR<1.5 for Rmax/Rmin=2 so the near neighbour interactions keep conceptually "representative" of a rock microstructure even though, I agree with you, it was not clearly explained. It is not easy to explain but I'll try anyway:

Lets consider a line of 3 particles 1, 2, 3 such that R1=2, R2=1 and R3=2. If you create bonds such as intR>1.5, then 1 and 3 will be bonded together. A bond will thus be created "over" 2.

In granular materials such as granite or sandstone (where only contacting grains interact one with each other), this is physically not acceptable. Then, if your idea is to discuss the influence of the microstructure of the medium on its overall behavior, this might be problematic. I must precise that this sort of long range interactions might nonetheless exist in some materials with exotic grain to grain interactions (e.g. Van Der Waals forces).

Now, even though this can be conceptually not exactly representative of the way grains interact in a geomaterial, it might nonetheless be an acceptable way to overcome some difficulties faced by classic bonded particle models to deal with certain types of cohesive materials (under the condition that no direct link is done between the numerical particles and the real grains). Another way to deal with such "non physically acceptable" interactions would be to use an enhanced contact law that would increased the degree of interlocking of grains (e.g. flat joint contact model in PFC, or moment contact law in YADE).

Luc

Thanks Luc and Robert, you solved my problem.

Thanks Robert Caulk, that solved my question.