Modal analysis question -- "free-free" condition 1

[271828]

Hello all. I have a question that I'm sure sounds very easy to many of you. Advice will be greatly appreciated.

Our team would like to modal test a 3,600 lb (6 ft x 8 ft, approximately) piece of masonry wall (orthotropic) and hopefully determine the flexural stiffness in both directions. We have already sawcut the wall and have it laying on the floor in the lab.

Our idea is to suspend the wall from its 4 corners using our overhead crane, with the idea of creating a condition that's as unrestrained as possible. The crane cables are probably about 3/8 in. diameter and are about 20 ft long. The testing program would be a simple quarter-point modal sweep with impulse hammer strikes at the middle of the wall. From that, we can plot the mode shapes for the different frequencies and hopefully solve for the stiffnesses.

I remember in experimental modal analysis class, we modal tested a small aluminum beam in the free condition by sitting it on soft foam. I also remember the teacher talking about suspending specimens using wire (small EA and large L) to create an unrestrained condition.

Would it be better to suspend the wall vertically from only two corners? Is there a better way to approach this problem than what we're planning so far? We do not have a small shaker, BTW.

Thanks in advance.
100,000e

 

[GregLocock]

I'd be very leery of using 4 cables as a constraint. You'd be OK if you just used one. The foam idea is much better - make sure that the six rigid body modes of the specimen on whatever suspension you use are nice and far away from the flexural modes of interest (having said that it may be that your 4 cable solution is good enough, if you have the test ready to go give it a try).

Usually we'd use 3/8 or half inch bungee cord for this sort of job if it was going to be a long one.

Next time you might try a modal on the wall in-situ, it should be possible to figure out the modes that are due to the foundaton, and still get your stiffness.

One of the more exciting discoveries from doing that is exactly how flexible a foundation is.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[271828]

Thanks Greg. What if we support the wall using two corners, so it hangs vertically? We're only interested in bending modes, so it seems like that would work ok.

We tested the wall in-situ, but couldn't isolate the wall's flexural stiffnesses. It doesn't just connect to a foundation. It starts on top of an elevated slab and then is supported at the top by a steel brace. We have frequencies and mode shape for the entire specimen, but there are too many variables (the wall-to-slab connection stiffness for example).

 

[GregLocock]

What are your cables made of? you'll get a nasty high frequency pitch (+1 -1) mode if they are steel cables.

Like I said, you'll learn far more from one hit than from any number of my notes!

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[271828]

The cables are steel. Keep in mind that we're interested in frequencies between about 10 Hz and 50 Hz for the tests. Does that alleviate the issue with the high frequency pitch?

 

[GregLocock]

C'mon, do the math! I don't know.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[271828]

LOL, I'm pretty sure the frequency will be outside the bandwidth we care about. Tomorrow, we're going to try to test it in the "hung by 2 points" vertical configuration.

 

[GregLocock]

Ah, another demonstration of "A month in the lab can often save a day in the library". Grins. Yes, hit it.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[MikeyP]

By all means do hit it, but can I suggest that the middle of the wall is not the best place to use as an excitation point. You won't excite the anti-symmetric modes very well. A few inches off centre should be fine.

M

--
Dr Michael F Platten

 

[sms]

Yes, hit it. I like the two point support better than the four point, but I don't have a good reason other than gut feeling.

"Why don't you knock it off with them negative waves? Why don't you dig how beautiful it is out here? Why don't you say something righteous and hopeful for a change?" Oddball, "Kelly's Heros" 1970

Please see FAQ731-376 for tips on how to make the best use of the Eng-Tips Forums.

 

[GregLocock]

The reason that two cables are better than 4 is that the 2 cables will affect the stiffness of the system in its wn plane, and provide no support in the out of plane direction, which is presumably the DOF of interest.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[271828]

Thanks everyone.

The tests, hanging from two points, went very well today. The first two theoretical mode shapes were about the same shape as the measured ones, but the order was flipped around. No big deal, though, because they were 68 Hz and 69 Hz--identical for our purposes. Rigid body modes were all down around 1 Hz or lower.

We were able to change the model in about 5 min. to get the theoretical to match the measurements, allowing us to back-calc. the elastic modulus. It was fully grouted, so EI was about the same in either direction.

We struck the wall away from the center in 3-4 places to excite some higher freq. modes, but these were of no interest to us. With the first two, we could calc. E, which was the goal.

Now we have another specimen to do the same with, but this one's much larger, like 20,000 lbf and is very orthotropic, like 3x stiffer in one direction than the other. We'll probably chop out a part and test it instead.

 

[GregLocock]

So what /was/ the pitch mode, out of interest?



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[271828]

Didn't even notice it. I think it must've been higher than 100 Hz which we used for the top of the bandwidth for most of the tests.

 

[GregLocock]

Oh well. Next time put an accelerometer pointing up near one wire and hit down at the other wire.

Then you'll see the pitch and bounce modes.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[271828]

Will do. Thanks for the tip.

 

[Vibac]

Hi 271828
Just a thought: Sounds like the model you updated from the measurements was a numerical one (FE?). Have you checked if an analytical formula for natural frequencies could be fitted to back-calculate E? I think Blevins has a set of formulaes for orthotropic plates.