Re: [Question #695058]: Bending beams mass computation

```Question #695058 on Yade changed:

Bruno Chareyre proposed the following answer:
You are welcome, nice to hear that someone is using it. :)

> In other words, what do I need to do in order to set the mass of a
body

Set the mass of its nodes. That's all. You could set mass to zero by
yourself for other objects  (or set density =0 and let preprocessing
compute the null mass). It should give same results.  So in fact you
don't even have to make it zero, you can just ignore it.

> PFacets seem to never receive a mass, but also the mass of the nodes
does not increase

Indeed, I don't think there is a helper function equivalent to "cylinderConnection()" for PFacet.
To be clear, cylinderConnection() would be better called "buildBeam()". A Pfacet equivalent would be "buildShell()" and it would assign thirds of facet mass to the nodes.
Currently the only option is to do that in your script. If you make it in a python function we can include it to source alongside the cylinderConnection().

> - What’s the difference between grid- and cylinder-connections.

I would say they are two alias for the same thing. Am I confused? have
seen one or the other in different scripts?

>- Why is it important to set a cohesive material for the nodes with
super high cohesive values?

Nodes material is used to deduce the properties of the interactions
between beam nodes (stiffness and strength), if cohesion is small you
can get get plastic deformation of the beam (in bending, shear,
traction). Not necessarily wrong, but if you want a simple linear
elastic response you set a super high cohesion.

Again, this is just preprocessing. You can always go loop on
interactions by yourself and change kn, ks, etc. regardless of which
material is assigned.

> How is the bending/mechanical behaviour in a node that connects 2
cylinders computed?

It isn't. Bending is defined for interactions, based on the differential
rotation of two nodes. It is not defined "per node". Same for the
connection between two facets.

> Can you recommend or provide resources for that topic?

Not really. You could check Effeindzourou PhD dissertation but I guess it's relatively close to the papers.
There is an old 2D version of the same thing described in [3] but I don't think it helps more if you already went through the 3D version.

Bruno