Simply Supported Steel I-Beam Test 2

[SrinivasAluri]

I am trying to test a steel I-Beam on simple supports, I've done a swept sine test using a shaker and accelerometer and calculated (Half-power method)the damping ratio to be 2.1%. I think is too high a number for an steel beam, my guess is it should be less than 1/2%. My question is could my setup distort (add) damping to the beam?? Resolution of swept sine is 0.05 Hz.

 

[GregLocock]

What frequency is the resonance?

2% is too high, try dangling the beam from an elastic cord and flick it with your thumbnail or a hammer - you should be able to get a good estimate of the damping ratio from the decay rate of the sinusoid.

There are several ways a bad setup can add damping, using a free-free test with an impact excitation will allow you to determine were the problem is.

Cheers

Greg Locock

 

[SrinivasAluri]

About 9 Hz is the resonance. I can't test free-free because I need the simple supports for adding TMD afterwards. I figured my supports could be iffy and so I am changing those to robust ones. Anything else I need to do??

 

[GregLocock]

What's TMD? This is a diagnostic test for your setup, not a change in the ultimate test.

Cheers

Greg Locock

 

[SrinivasAluri]

Oops! TMD stands for tuned mass damper.Thanks for the tip. I'll post the progress of the test.

 

[MikeyP]

Simple supports are just about the hardest "textbook" boundary conditions to emulate in practice. You can get good approximations to clamped conditions in many cases and excellent approximations to free conditions. The problem with simple supports is that you need very robust end supports to hold the end at a fixed point, but the end must also be free to rotate. This usually means some sort of hinge mechanism where freeplay can be a problem. You may have to use an axle on bearings at the ends to get a decent approximation to a simple support.

I have a more general comment about your test. I assume that you are trying to find the best place on the beam to position your TMD. That optimium position will be determined by the mode shapes of the beam. The mode shapes in turn will be determined by the boundary conditions. So you should test the beam in the boundary configuration in which it is to be used. Is this beam going to have simple support end conditions in practice?

M

--
Dr Michael F Platten

 

[SrinivasAluri]

I agree with you Dr.Platten, free-free test is the simplest. The actual beam I am trying to simulate has simple support ends.

 

[sreid]

As stated by others above, solid metal beams have very low damping ratios in a free-free condition, they ring like a bell. Any supports can add large amounts of damping. Also you might try sweeping the excitation slower. It takes considerable time to build up amplitude in a low damped system and if the frequency sweep is to fast the resonate peak will not reach full amplitude.

 

[SrinivasAluri]

I was thinking it was the other way round, heavily damped systems need slow sweeps not the lightly damped ones? Am I wrong??

 

[MikeyP]

No, or rather, yes you are wrong, sreid is right. This is the reason I don't like using sine sweeps for testing. In a sine sweep test the vibrations never reach steady state. The response therefore always contains some transient component. In a lightly damped system, these transients take a long time to die away. In a highly damped system they die away much more quickly. If you are sweeping too fast then the peak may well be in the wrong place or wider than you expect and hence your damping estimates will be inaccuarate.

Are you using a current controlled amplifier to drive the shaker (one which provides an output current proportional to the input voltage)? If not, then this could also distort your response curven due to a nasty little phenomenon called "force drop-out".

Usung half power points to estimate damping applies specifically to the FRF, which you are not measuring. Greg's rough and ready impact approach will probably give you a better estimate than your half power approach.

A better method is "sine dwell" which is marginally more sophisticated than the impact approach in that it allows you to estimate the damping of individual modes rather than the structure as a whole. It also doesn't require any more equipment than you already have. It should work well on your structure because the natural frequencies are well separated.

1. Put a sine wave through your shaker as close as you can to the natural frequency.
2. Apply it for long enough so that the structure reaches steady state.
3. Switch it off and measure the decaying reponse.
4. Use log-dec or similar to estimate the slope of the log of the decay.

Not so many years ago, we used to estimate the slope by offering a protractor up to the 'scope screen. Fortunately things are a little more sophisticated now.

M



--
Dr Michael F Platten

 

[SrinivasAluri]

Thanks Mikey for the detailed clarification on the sweptsine testing, the sweep time at each frequency is 5 secs (is it high or low?). I was reading a book on Vibration testing and I realized that I was using the amplifier in voltage mode (which is what Mikey is saying). I did get the same result when I used the exponential decay method, which now I think is due to both bad support conditions and using the amplifier in voltage mode. Also, there is some electrical noise in my pure sine signals, which the book says is due to coiling of accel. cables because they are too long!!! Hopefully all these changes will improve the accuracy!

 

[SrinivasAluri]

Thanks guys! We finally built the robust setup, used the amplifier in current mode, unrolled the coils and got a damping of 0.5%.