Nastran / Frequency Response / Free-Free

There are examples of using SOL 108 and SOL 111 to create a frequency response. In Siemens Simcenter (NX Nastran), the examples focus on a tower with a seismic mass. The model drives the mass with an acceleration, attaching the mass to the structure using rigid elements.

This creates a model that adds (?) stiffness at the mass attachment points.

What if I want to simulate the frequency response function (FRF) of a structure that is unconstrained?

The test I conducted was to hang my part on a wire and tap (laterally) it with an impulse hammer.

This might be equivalent to constraining the support point as:

• Fixed in DOF's 1, 2 (Tx and Ty)

I could apply a lateral acceleration (in the direction of the hammer tap).

Is this the conventional way to produce the FRF? If not, what is?

If your structure was literally hung on a piece of (plastic covered) wire then it was constrained vertically, at that point, by the wire.

A better method is to hang it from a soft bungee cord, so its natural frequency of vibration as a "mass on a spring" moving vertically is a factor of 10 (or more) lower than the frequencies you are interested in measuring.

But to answer the question, to find a FRF for an unconstrained structure in Nastran, you just run the analysis with no constraints and no other "tweaks". So long as you don't try to calculate the response at a frequency of 0 Hz (which obviously doesn't mean anything physically), the analysis should "just work".

Modelling your "stiff" wire support correctly is actually quite difficult, because if you think about it, the structure can behave like a compound pendulum consisting of a rigid wire and a rigid mass, with different vibration modes in different planes because of the different rotational inertia properties of the mass in different planes. Those responses are not "linear" behaviour, because they depend on the stress distribution in the model caused by its self weight - the most significant effect being the tension in the string.