How to identify the natural frequency for the phase/magnitude? 2

[0922]

I did a bump test. In the magnitude chart in FRF, there are few peaks, but none of them corresponds to +90degree in the phase chart, but there is one points to -90 degree.

Can I say the peaks are natural frequency anyways? please help

 

[0922]

Also, I found the driving point has always max movement, why is that?

 

[GregLocock]

In real life systems often have resonances that are not at exact quadrature. If you look at the Nyquist plot you'll see why.

If the driving point has the highest response then you may have problems with technique or setup or analysis.

Cheers

Greg Locock


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

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

 

[0922]

Great to hear reply. A littl more detail. I made a cylinder model which consists of three layers' circles. The accelerometer was fixed to one point and other points were hammered to get the whole system's FRF, and the driving point showed the max displacement in all cases. How can I fix this?

 

[GregLocock]

"If the driving point has the highest response then you may have problems with technique or setup or analysis. "

Hard to narrow itdown much. Can you upload your reciprocity test?

Cheers

Greg Locock


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

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

 

[electricpete]

I don’t do experimental modal analysis. Just a little curious....

I’m understanding right:
driving point impedance = response / excitation....at location of excitation
other impedance = response / excitation.... at other location

Then, why would be be suprised to see driving point impedance > other impedance... i.e. max response at the location of the excitation ?


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

 

[0922]

Attache is model, results and one of charts. The driving point doesn't effect modes, but does effect the magnitude in my test. We also want to know where is max movement againist some frequency.
Thanks.

 

[0922]

Model is as linked.

 

[0922]

Model

http://files.engineering.com/getfile.aspx?folder=37caa76b-c50a-4c7a-9d34-81e4eb461594&file=model.PNG

 

[0922]

Results

http://files.engineering.com/getfil...c505f&file=result-magnitude_accePerNewton.PNG

one of charts

http://files.engineering.com/getfile.aspx?folder=78ff0c2d-38b9-4aa1-bc65-88e622840b86&file=chart.PNG

 

[electricpete]

The FRF phase seems to show a tendency to continuosly decrease, except for abrupt phase increases. Most of the abrupt phase increases are wrap-around of scale. Let’s look at some of the other ones:

One abrupt phase increase at 420hz occurs at location where a zero is very evident from the magnitude. That’s exactly what we’d expect (assuming convention that positive phase angle corresponds to response leading excitation)

Another phase change one as 220hz doesn’t make any sense to me.

I’m trying to think of what a linearly decreasing phase means. One thing that comes to mind from Fourier theory is a time delay (delay a signal by T0 by multiplying it by exp(-i*T0).

At any rate, the change in phase does not seem related to the peaks in magnitude as I would expect (would expect relatively distinct decreases in phase at locations of the peaks). Maybe this linear phase increase obscures the underlying behavior somehow.

Just my rambles fwiw (not being a big modal analysis guy myself). If Greg or others have some rambles, I’m sure they’ll be worth more.


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

 

[GregLocock]

Your hammer is too big and you probably have setup problems. Check reciprocity first if that doesn't work nothing else will.

Cheers

Greg Locock


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

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

 

[0922]

Thanks for Greg's comments. My hammer has 1 lbs weight. Lots of people use 3 lbs hammer in this area, but this small one works.

Just wanted to tell what I got. For the test mentioned above, the reciprocity was so bad and not consistent at all. I moved the test points from "the bad spot" to more rigid spot, and it gave me very good results. No issue any more.

I am thinking "the bad spot". Those area are filled with epoxy material and not consistent at all. I guess the reason might be not only the high damping but also the non-consistent filling or say non-linear problem.

Again, thanks Greg

 

[0922]

One more thing forget to mention. When I have to do modal analysis on the highly damped structure, perhaps I may move the hammer along with the specimen accelerometer in the same time. I'll take some test to verify this thought.

 

[GregLocock]

Nothing wrong with that but make sure your matrix of driving point and response point locations is sufficiently complete. eg if you've 1x 1x and 2y 2y make sure you've got 1x 2y or 2y 1x in there as well.

Cheers

Greg Locock


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

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