Incorporating Vibration Mounts in FEA 1

[CaptainCrunch]

Hello All,

Has anyone done any work on incorporating the vibration mounts into Finite element analysis (ABAQUS or NASTRAN)?

the only way I could think of it, is to measure the spectral transmissibility of the mount by itself, then define a material damping of the mount material which is a function of frequnecy. But it isnt clear to me what damping values will lead to a *specific* transmissibility at a specific frequency.

Just wondering if anybody has been down this road before, and would like to share their experience.

Thanks,

George Vandyke

 

[tomirvine]

The primary effect of isolator mounts is to lower the fundamental frequency of the system. Ideally, the isolators also render the system as a single-degree-of-freedom system.

The isolation frequency should be well below the dominant excitation frequency. Damping is a secondary effect if this condition is met.

The isolation frequency should also be well below the component's fundamental frequency in the hardmounted configration.

I have modeled isolated systems in Femap/Nastran. I modeled the isolators as springs. The property card has stiffness and damping options. I specified a stiffness value for each spring but left the property damping as zero.

I then specified a modal damping value as a function of frequency. The modal damping value is for the system as a whole.

Tom Irvine

http://www.vibrationdata.com

 

[GregLocock]

A case in point is say the engine mounts in a car. Road excitation of the engine is a very important mechanism, and it excites the primary bounce mode of the engine/mount system.

We no longer build NVH models in Nastran. We used to but the spring/damping coefficient model is to restrictive. Nowadays we model the big lumps (engine, gerbox, body) in Nastran, but assemble them in programs like VSIGN and others made by Pius, where we can use non linear frequency dependent characteristics for isolators. This also makes it easy to add in measured characteristics, and excitations.

Cheers

Greg Locock

 

[CaptainCrunch]

Greg,

How do you determine / test / define the below term you used? Could you give more detail?

"non linear frequency dependent characteristics for isolators"

George Vandyke

 

[GregLocock]

Test. You mount the engine mount on a hydraulic shaker, apply the correct preloads, and then map the force vs displacement characteristics for all relevant frequencies and vibration amplitudes.

It gets to be even more fun when you have a suspension bush which has to have known characteristics under different preloads as well.

Even a plain ordinary rubber bush will have frequency dependent stiffness and damping, and that will change as you change the preload and the amplitude of vibration. I am keen to measure the effect of loads in other directions to the one being measured as well, but our Test lab is rather overworked.

Incidentally a spring and damper model does not match the dynamic characteristics of the bushes I use, since they tend to have a roughly constant, but increasing, stiffness, and a phase angle that remains almost constant, with increasing frequency. This makes selecting a damping coefficient rather tricky for broadband excitations. Cheers

Greg Locock