FEA with grav acceleration and mass point

[Henry26]

Hello everyone,

I'm performing a FEA with gravity acceleration as load case and a point of mass concentration located at CG. I linked this point of mass with a rigid link in all inner surface of the componente. My doubt is: Is it correct link this point of mass only in inner surface? Or should I link that with others surfaces, like outer surface.

Thanks in advance!

 

[rb1957]

I'd model the structure with the correct density and apply a body force (acceleration).

If there's a lump, model as an RBE to it's CG and attach to the rest of the world at the appropriate fasteners.

another day in paradise, or is paradise one day closer ?

 

[Henry26]

Hello rb1957,

Thank you for your answer. I agree with you about using the structure with all the respectivity density. But the reason to use the point of mass is to simplify the analysis. That's exactly what I've done. I've used a RBE3 linked from the inner surface to the CG with point of mass. But I'm not sure if use only the inner surface of the structure is the correct way to perform this kind of analysis.

Thank you.

 

[rb1957]

what's so hard about adding density to the material properties ?

You're point loading the structure with it's distributed inertial load. I don't think anyone would argue inner- or outer- or both surface nodes.

the correct way to perform this analysis is via the material density. everything else is a rough approximation (depending on how many nodes you distribute the weight over). "fussing" over inner- or outer- or both surfaces is like "counting the number of angels dancing on the point of a needle".

another day in paradise, or is paradise one day closer ?

 

[Mustaine3]

I linked this point of mass with a rigid link in all inner surface of the componente.

Then you've made those regions rigid. Do you really think, that this is appropriate to get useful results?

 

[rb1957]

I thought an RBE3 didn't add stiffness to the model ?

another day in paradise, or is paradise one day closer ?

 

[Mustaine3]

I just saw that you wrote 'rigid link' in the initial post.

A RBE3 has to be respected by the solver and could alter the behavior of the part. It also extends the bandwidth of the system, so that more memory is needed. The more nodes you have in the RBE3, the worse it gets.

That the coupling of specific regions to a mass point is different than having mass in the whole part was already mentioned.

 

[FSB1]

you can tune the mass of your model by increasing the density

or you can add a non-structural mass distributed to all elements of the model, the result shall be the same