Modal Impact Test Question

[PLMKY]

As part of a research project I will be load testing a beam in my school lab. I have been asked to try and determine the damping of the beam as part of this project and I was considering conducting a modal impact test. I am not very familiar with structural dynamics and wanted the opinion of those more experienced. The span of the simply supported beam is about 3.3 metres and it is 300mm in depth.

I do not have the funding to rent or buy a proper modal impact hammer. Therefore, if I were to conduct this test I would be striking the beam with a regular hammer while ONLY measuring the acceleration of the beam. I recognize that you should have a load cell on the hammer head in order to determine the impact duration and magnitude for proper modal analysis. However I would like to know how much inaccuracy would I be introducing if I were to only measure the output accelerations? I realize I would not be able to produce a coherence plot in order to check the quality of my data. However, I am primarily only concerned with accurate identification of the first mode of vibration and I would be looking to verify the first natural frequency and produce a corresponding estimate of the damping ratio. I am thinking of placing the accelerometer at mid-span of the beam and I would strike with the hammer only at mid-span. This would be done to ensure that data for the first mode shape is the most accurate. I would conduct this test maybe 5-10 times and take the average results in order to help reduce the error due to noise and inaccuracies from differing pulse durations.

Any comments you can provide would be greatly appreciated. I am not familiar with this method of analysis and I am learning as I go. I realize that my data using this method may not be the most accurate but I would like to provide reasonable values for minimal cost as this is simply an addition to the main purpose of my research. Any insight, resources and comments you can provide would be greatly appreciated!

 

[GregLocock]

If you are only interested in the damping and frequency then a plain ordinary hammer is fine. Use logarithmic decrement for the damping and the obvious for frequency.

Cheers

Greg Locock


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

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

 

[Vibac]

Adding to GregLocock's comment, I suggest reading this nice paper by B&K, which compares different methods:

https://www.google.com/url?q=https:...FjAAegQIBBAB&usg=AOvVaw3B8rY8z8pzzN3uDnV-vxUD

 

[PLMKY]

Just to clarify my understanding of the subject:

In a typical modal impact test you strike the object with a hammer and measure both the input and output responses from the hammer and object respectively. Since the impact is not a pure impulse it is treated as a harmonic load and frequency analysis is used to develop the response spectrum for identification of the natural frequencies, modal shapes, and damping in the system.

In my proposed setup since I will only be measuring the output acceleration of the system AND the impact load will be applied only at mid-span to initiate the first mode of vibration, I am assuming the application of a pure impulse to a system with only a single-degree of freedom (SDOF). When a pure impulse is applied to a SDOF system, the system is put into free vibration and I can use the logarithmic decrement method to determine damping and I can count the cycles to determine the natural frequency.

I imagine since my beam is truly a continuous system in addition to the fact that I am not applying a pure impulse, there maybe be excitations for different mode shapes however this will be small and blend into the noise I will be capturing with my data?

Thanks again for the help. Much appreciated.

 

[scheah]

PLMKY,

Logarithmic decrement method works nicely for SDOF textbook problems. For your beam test with multiple modes getting excited simultaneously, a half power method should work better. See this simple writeup on how to apply it.

If the cross section of your beam has similar width and thickness, there is likely closely spaced modes. This is not easily distinguished in a output only measurement.

Talk to your professor about your challenges. I would also try reaching out to the local equipment representative to see if you could borrow an impact hammer. No harm asking.


Kind regards,
Jason

 

[GregLocock]

We've already agreed that it isn't ideal, so there's no point worrying about details. In your case most of your response will be the first mode.

Incidentally don't average the time waveforms together before analysing them, analyse each one and then throw out any outliers and average the remaining results.

One way of getting a clean input signal is to use a relaxation method. So tie a rope around the beam, tension it, cut it.

Cheers

Greg Locock


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

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

 

[scheah]

it isn't ideal, so there's no point worrying about details.

I like your attitude Greg

Depending on the tip of the hammer, more than 1 mode can be excited. An example time domain output in this paper on a simple reinforced concrete beam on Figure 3 shows log decrement method is difficult to apply.

Kind regards,
Jason

 

[GregLocock]

Didn't they suggest a method to resolve it? Low pass filter. Or as you say, the half power method. Which is pretty much the same in the end.


Cheers

Greg Locock


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

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