Cylinders: errors in the mass and inertia matrix calculation ?
Hi all,
I am trying to perform simulation of the flow of cylindrical particles down an inclined plane. I succeeded to set a simulation and created a new contact law to modify the existing contact law (Law2_GridCoGridCoGeom_FrictPhys_CundallStrack()) to account for dissipation at contact (spring-dashpot type contact law). While trying to validate the new contact law, I figured out what seems to me to be two errors in the way the mass and the inertia matrix of the cylinders are evaluated in YADE :
- The mass of the cylinder takes only into account the cylindrical volume defined by the grid connection but not the volume of the two ending half spheres/grid nodes. Meanwhile, the grid nodes volume is taking into account in the contact detection. Maybe this is due to the fact cylinder are mostly used as chained cylinder ?
- The inertia matrix of the cylinder is calculated in a way that seems unaccurate, considering an equivalent sphere instead of a cylindrical shape. When considering a clumped cylinder, the inertia matrix computed does not correspond to the classical expression of the inertia matrix of a cylinder. Taking into account the two grid nodes inside the expression of the inertia matrix does not solve the problem. Is there a reason for which it is computed this way ?
I wrote a note to detail the problems (attached) and made a minimal script to illustrate the fact that it is indeed calculated in YADE as I wrote in the document (also attached). I want to know whether there are some particular reasons I did not identify for this two points (and then I will just modify by hand the mass and the inertia of the cylinders), or if there is a real mistake (in particular for the inertia matrix) which should be fixed in the code. Could anyone familiar with the use of cylinders in YADE have a look at the note to give me a feedback ?
Regards,
Raphaël