acceleration recorder of massless node

[claytp]

I have a building model with the seismic masses lumped at the beam-to-column joint nodes. All other nodes are massless. When I record the accelerations at the beam-column joints (that have mass) I get reasonable floor accelerations. In the model, the beams are made up of multiple nonlinearBeamColumn elements. When I record the acceleration of one of the massless nodes along the interior of the beam, I get very large values that grow with time during the earthquake simulation.

My question: how is it calculating the acceleration response at massless nodes?

I would have first assumed I would get nodal accelerations within the beam similar to nodal accelerations at the lumped mass nodes. These nodes would still have the same acceleration, but since they have zero mass, there are no inertial forces acting on the node.

However, since the nodal accelerations of massless nodes give numerically infinite values that grow over time, it looks to me as if the nodal acceleration is being calculated by taking the unbalanced force at a node and dividing by a negligible mass. Even with this though, it seems like there should be some divide by zero error since the node actually has zero mass.

[yigitozcelik]

i am not sure but you can try to assign negligible (1.e-20) masses at non-mass joints.
You may know similar problem happens with eigen command,
as i said, not sure, but may figure it out

[fmk]

look at the theory for Newmark

[url]
http://opensees.berkeley.edu/wiki/index ... ark_Method
[/url]

mass less and mass dof are treated the same

do you have damping?

[claytp]

Yes I do have Rayleigh damping (2% in the 1st and 4th modes for a 9 story building).

Since the newmark method does treat massless and massed dof's the same, I would have thought all nodes (at least within the same level) would have similar accelerations since accelerations are used in the U(t+dt) expression and U should not vary that drastically along a floor beam. Then the inertial effects would be governed by the presence or lack of mass. But this doesn't seem to be the case...

[mnazari]

I have the same problem. How did you solve it?