Rule-of-thumb modal test

[Spirit]

Hi pals, any of you has previous experiences in performing impulse modal testing WITHOUT using an instrumented hammer, interpreting the modes directly from the accelerometers signals? I guess this is a kind of 'extreme sport' in modal testing, but that's it...
Thanks, bye


Spirit
'Ability is 10% inspiration and 90% perspiration.'

 

[vorwald]

I've done modal testing with and without recording the input force. A rap test is useful to identify the first 3 to 5 frequencies. Frequently, this is all that is needed, for example on shaft dynamics. Also, you can estimate the damping of the first mode using log decrement on the time history data.

A second approach to input the force without measuring it is to deflect the structure with a string and then cut the string.

I'm not sure what the question is, can you reword it?

 

[MikeyP]

The trouble with vorwald's approach is that you do not get a scaled modal model (i.e. you can't get the modal mass) but if you just want vague estimates of the mode shapes frequencies and dampings then it should be ok as long as the modes are lightly damped and well separated in frequency. It won't be accurate though.

There are ways of getting better estimates of the modal parameters (although not the modal mass) using "operational modal analysis" but I think I'm right in saying that these only work with statistically stationary signanls, eg. broadband random exitation.

M

 

[GregLocock]

The old SD 380 had a really cute trick. You could waterfall the magnitude of each spectrum from each impact test, so you could just tap a simple system along its length and build up pictures of the first few mode shapes, all on screen at once. It was much better if you had a force signal as well, then you could use the imaginary part of the FRF and get pretty mode shapes.

Also people use to do this a lot with acoustics - clapping or bursting a balloon to get the first few room modes.

So, yes you can do it, if you've got the time to waste. The rest of us use a 2 channel analyser.

Cheers

Greg Locock

 

[Franko]

For some structures, rap tests are a quick method to confirm hand calculations or FEA analysis, or if there is interpolation available from similar structures with known frequencies corresponding to mode shapes. (Modal mass is not required for every analysis - resonance and/or analytical correlations are the overwhelming consideration). An example, besides lifting a rotor with nylon straps to obtain free-free modes, is the first four modes for an axial compressor or steam turbine blade - with proper fixation at the blade root:
a) striking perpendicular to the neautral axis for a particular direction will give the the fundamental tangential and axial modes. Striking the blade at various points along the tip and only measuring response can show the node point when you strike there for the first torsional mode. This ferrets out the second tangential mode which is sometimes fairly close to the first torsional. With a two-channel analyzer and an instrumented hammer, frequency response function amplitudes can be measured without doing a full modal - to save time. I imagine others could do the same and save time for other types of structures. Log decrement values as one respondant suggested are almost always meaningless from light rap tests. We don't even document them when using transfer functions from 2-channel analyzers.

 

[MachineryWatch]

You have received a lot of good responses. Of course we are all answering based on our own backgrounds and experience, that is expected. You asked for our experience, right? I think in order to really answer intelligently we need to know a little more about your application. Are you really trying to do a modal analysis, or are you just trying to make a determination of whether or not you have a resonant condition and, if so, find a "quick and dirty" solution to a problem in the field.

Most of my work is machinery diagnostics in the field where the budget is extremely limited. The main goal is to identify the source of the problem as quickly as possible and correct it. "Modal" analysis is usually out of the question (often we can't even shut the machine down!). We have to use many different techniques to make a determination of the presence or absence of a resonant condition. Most people involved in the use of true modal analysis would consider what I do to be incredably inaccurate, but it is hard to argue with results that solve problems.

What is your application?

Skip Hartman

http://www.machinerywatch.com

 

[Spirit]

I agree with you Skip, I received a lot of good responses from people who accept to share their experiences and background to help colleagues and learn something. Yes, as I wrote in my first post I am trying to perform a modal analysis, probably it is not so important to specify exactly my application, since I see I received more than intelligent answers from colleagues that have considered sufficient what I wrote in my original question. We mostly work in private companies, and it is not always possible to share with our colleagues what our firms are working on. Luckily it is not so important to have these details to receive good advice from people who agree to share common experiences to grow up. I hope this is your commitment, too.

Cheers


Spirit
'Ability is 10% inspiration and 90% perspiration.'