Nodal velocity as boundary condition

[matparta]

Hello
I am interested in an harmonic analysis (frequency response) of a structure made of a mass connected to a frame. The mass can be modeled as a point mass, connected with rigid constrains (or distributed coupling) to the frame, which is modeled by shell elements.
I want to apply a velocity on the mass as a nodal boundary condition - in other words, I want to simulate a mass moving with given m/s at given frequency(ies) and I want to see the vibration velocity in a certain part of the frame.
Now, the FE softwares that I have checked do not have the possibility to apply such as boundary conditions (Ansys and Actran - Actran seems to be able to do it but it is cumbersome). I can only apply nodal displacement boundary conditions or nodal loads.
Why is that? How would you approach the problem?

I thought about converting the velocity to an acceleration and by knowing the mass of structure connected to a frame, convert that to a point load applied as a nodal load.
Thanks

 

[GregLocock]

why not turn it into a displacement at each frequency w using v=j*w*d?

It almost sounds like a forced response than a FR.

 

[BlasMolero]

Dear Matparta,
When the excitation is defined in amplitude of displacement or velocity in the frequency range the correct approach is to convert the excitation to acceleration and apply to the base of the frame structure using an RBE2 spider rigid element, not to the point mass: in the point mass is where you will read the response and plot vs. frequency of any response value you like : displacement, velocity or acceleration. This is the typical approach I do when solving dynamic modal frequency harmonic response analysis (SOL111) using Simcenter FEMAP with NASTRAN.
Best regards,
Blas.

 

[matparta]

Thanks.
I have run a few tests but I am confused by getting different results in respect to magnitudes.
The model is simple and is made up of a U-beam modeled with shell elements, on top of it at a distance there is a point mass, connected to the beam via rigid elements. The beam is constrained at 4 edges. The mass shall be prescribed to move with velocity v (I am using Ansys and I use d).

These are the tests I have run:

1. point mass = 10 kg; displacement boundary condition d=v/jw applied to the point mass.
2. point mass = 1e-6 kg (negligible); point load F = jwv * 10 kg.

I should get comparable results, but I do not, they differ by a order of magnitude, but the pattern (color fringe in the plot) is the same.

1. The maximum displacement in the result file is indeed d=v/jw and it is at the point mass location and at the nodes at which the rigid constraints are attached (as expected!). The reaction forces are constrained nodes are zero (why is that? OK, there is no actual force load in the model but that harmonic displacement is causing a harmonic acceleration which times the point mass is a load).
2. The maximum displacement in the result file is at the same locations as in (1) but it has a smaller magnitude by circa a factor 20. Reaction forces at constrain add up to F.

The models should be equivalent and besides any difference due to how the FEM routines handle models with different loads and or BC.