measirung excitation forces

[white123]

Hello,

I would like to ask if someone has an experiences how to measure excitation forces, for example: Imagine that you have an electric engine mounted on the structure and you need to measure the excitation forces from engine applied on the structure - what will be the best procedure? how to get "clean" forces (not including structure eigen-frequency responses)?

thank you for your answer!

 

[IRstuff]

As a general rule, I don't think that you do these types of measurements, in situ; there's probably too many interactions to get an uncluttered picture. What I have heard is that people often suspend such things in the air.

TTFN
faq731-376

 

[amanuensis]

First of all, a little bit of theoretical background... If the frame is light compared to the engine, then the coupling force will be important, will have a high value.

There are three axis dynamical force sensors which can be fixed between the both structures (see Kistler for instance)

But if you can not afford to buy these sensors, then you can asses the coupling using substructuring method.
In this case, the coupling force is given by
F=V^2/(Y1+Y2) where Y is the mobility i.e. FRF.
So you need to have at least an impact hammer to do this.

 

[GregLocock]

We cut the structure apart and measure the loads with a load cell, or strain gauge the structure and measure the strains, or we measure the input impedance of the structure and the vibration on it, or we measure the properties of an isolator and measure the difference in vibration on each side.

A relatively easy approach would be to add a known imbalance to your motor and measure the structure's vibration, and then compare it with the original value, if you are worried about balance forces.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Vibac]

Amenuensis,
Using your mobility and velocity based approach, wouldn't you miss the energy corresponding to rotation around the three axes? I wonder in which cases this might still be a fair approximation?

 

[amanuensis]

You're absolutely right, vibac. Rotations are not taken into account just because it's not possible to measure them! Their influences are supposed to be negligible...

It's a technological problem that obliges us to do this assumption.

 

[Strong]

I don't know about the modeling issues, but I would guess that neglecting torsional forces/vibrations would be unwise for an electric "engine".

Walt

 

[amanuensis]

It can be reasonably conceived that torsional vibration is strongly coupled with both longitudinal and tranversal vibration
Under this hypothese,measurement of translation acceleration includes the effect of rotation. The torsional vibration can be seen as hidden variable...

 

[white123]

Hello, thank you all for your answers!

For better imagination I created a simple image,. the motor (black) is mounted on bigger plastic part (red) in four locations.

The problem is that we need a "clean" forces from measuring (not including red structures eigen resonances) but we don't know how to measure that.

Let's say that our budget is unlimited so If anyone can tell us how to measure this problem and what equipment we should use? this is the picture (hand drawn in paint)

 

[amanuensis]

Did you draw your picture with your motor operating and your PC on the structure???
Cause it's very trembling...
Anyway, your issue is just the basic source-isolator-receiver problem.

It seems that you assume that the isolator is a structural element with distributed mass and elasticity.

You should then consider your model in terms of impedance(Z) and determine the Force Transmissibility: Fr/F=1/(1+Zs/ZI+Zs/Zr)

 

[GregLocock]

If money were no problem I would install 6 dof load cells in each mount.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[amanuensis]

Hi Greg,
Don't you rather mean 3 dof instead of 6 dof.. (but maybe you wanted to make a joke and then it was very funny)

 

[GregLocock]

No I meant 6.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[amanuensis]

Well, it is very mysterious...
Could you please reveal your scoop?

 

[GregLocock]

X Y Z RX RY RZ

obviously

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[electricpete]

i.e.: translation along 3 axes and rotation about 3 axes

=====================================
(2B)+(2B)' ?

 

[amanuensis]

thanks the great physicists. Everyone knows what are the 6dof...
The only problem is that Rx Ry and Rz are not measurable. All the dificulty is here.
I thought that greg knew a sensor able to measure the rotations.

 

[GregLocock]

I certainly do. Kistler and others make them. I happen to have 8 of them in the electrical lab for calibration at the moment. That's about $2 million worth.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?