Mode matching (measured modes and modes obtained from FEA)

[kajalschopra]

Hi,I have been reading a text on "Finite Element model updating". This question is not related to the subject of "Finite Element Model Updating" but rather to the subject of "Structural Dynamics" and experimental modal analysis.

My questions is:

1)Let us say I have carried out a free vibration analysis of a system (let us say single storey frame)

2) The text says that,

"N number of observed modes in dynamic tests may not necessarily be the N lowest-frequency modes in practise. In other words, some lower modes might not be detected. For example, some torsional modes are not excited. Further, in case there is damage in the structure, the order of modes may shift because local loss of stiffness from damage may affect some modal frequencies more than others"

I do not understand this concept of mode matching. Can someone put some physical insight. Let us say I have a simply supported beam. I get 5 modes following a modal analysis.

Why is that if I experimentally obtain the first 5 modes, then the numerically obtained modes by FEA not match the measured modes (With respect to the above stated by the author)

With respects
Kajal

 

[Dr.D]

Let's agree to limit the discussion to your simply supported beam with 5 modes found by modal analysis.

You know from theory where the node points should be in that simply supported beam and where the loops should be. When you look at the experimental data, does it show the nodes at the expected locations?

Imagine for a moment that your experimental data shows that the lowest mode measured has a node at mid-span. This is not where the lowest mode should be from a theoretical perspective, but it would suggest that the lowest mode measured is actually the second mode. This is where the idea of comparing theory and experiment can safe guard your understanding of the measured data.

 

[xxChrisxx]

My mathematical theory on this is not the sharpest. However I have done a fair amount of practical modal testing.

From an quick reading of point 2 it just appears to be a fairly general warning that idealised assumptions don't always reflect reality.

You took the case of a simply supported beam. I assume it was 2D beam, or at least a a case where your force is acting in the plane of bending.

Where in reality are you going to experiment on a simply supported 2D beam?
Is it really simply supported?
Are you always going to apply exitation exactly in the plane of bending?

 

[kajalschopra]

Thanks you very much for the responses.

Just to give you a background, I'm considering the topic that I'm going to update my FE model using the experimental data. [Yes, this is just a fundamental understanding to get hold of the concept. Yes, in practise one may not do the measurement of a simply supported beam or one may not have a simply supported beam in a practical world.]

My theoretical (or finite element) mode shapes are as follows:

You said:

Imagine for a moment that your experimental data shows that the lowest mode measured has a node at mid-span. This is not where the lowest mode should be from a theoretical perspective, but it would suggest that the lowest mode measured is actually the second mode

Do you mean to say that the measurement HAS MISSED THE FIRST MODE OR DO YOU MEAN TO SAY THAT THE FIRST MODE DOES NOT EXIST IN REALITY BUT ONLY IN THEORY? I believe the author means that the lower mode might not be excited (that is, the experiment has missed it)

N number of observed modes in dynamic tests may not necessarily be the N lowest-frequency modes in practise. In other words, some lower modes might not be detected. For example, some torsional modes are not excited

Thanks again,
Kajal

 

[Nidum]

I don't think you have clearly identified what you are trying to do .

If you get a mismatch between test results and calculated results then you go around the loop again as many times as necessary and find out what's gone wrong .

Basically just the engineers version of scientific method .

 

[xxChrisxx]

Do you mean to say that the easurement HAS MISSED THE FIRST MODE OR DO YOU MEAN TO SAY THAT THE FIRST MODE DOES NOT EXIST IN REALITY BUT ONLY IN THEORY? I believe the author means that the lower mode might not be excited (that is, the experiment has missed

It exists in reality and realistically you'd always pick it up.

Consider that you have to physically excite the system to make a measurement, the best place to do this is at an antinode.

For mode 1: let's say you chose to excite the system near one of the nodes. You may not get enough into excite the big first mode.

 

[Randy Beikmann]

In finding modes on physical systems, it is often necessary to excite the structure at several locations and/or more than one direction. Any single shaker input applied to a 3-D structure is likely to fail to excite 1 or more modes sufficiently. This is because any particular DOF is likely to be at or near a node for one or more modes. Of course it also depends on what frequency range of modes you are interested in.

When more than one excitation is used, you can do them one at a time, but there are techniques to use multiple inputs at once.