Modal Parameter Extraction - SSI VI

[bk.]

Hello all,

I have a few questions about the SSI VI that comes in the modal parameter extraction download. 

To give some background:

 - I am running Labview 2011 with the advanced signal processing toolkit and am using the SSI modal parameter VI as previously mentioned.

 - What I am trying to do is determine the mode shapes of a bridge model I have developed in the structural analysis program SAP2000. 

 - I am obtaining simulated acceleration response data at a sampling rate of 200Hz from SAP and am inputting this data into Labview, then running it through the SSI VI

Due to the limited documentation and instructions about the modal parameter extraction VI's, I am unsure if the output from the SSI VI is correct.

What I have found so far: 

- The mode number sets the number of modes to be found.  When I set this value to an integer greater than 0 a certain number of mode shapes are determined based on the percentage of noise subspace added. 

- If the number of modes to be found is set to 0, a certain  number of mode shapes is determined regardless of the percentage of noise subspace added.

- All of the modes I have determined so far from the VI are complex modes, which seems to be correct regarding previous operational modal analysis.

- Also, if I change my sampling rate from 200 Hz to 2000 Hz, which would seem to make the results more accurate, it raises the magnitude of the frequencies determined by a power of 10.  So for a 200 Hz sampling rate frequencies ranged from 4 Hz to 100 Hz and with a sampling rate of 2000 Hz, frequencies were in the 400 Hz to 900 Hz range.

My Questions:

1) Is it correct to set the number of modes to be found to 0 so that the noise subspace does not affect the results?  This way still determines a set number of mode shapes for the inputted data.

2) If this is incorrect, what is a standard noise subspace percentage to assume for analysis?  The program defaults to 50%.

3) Why does the sampling rate change the magnitude of the determined modal frequencies?

4) The residual vector is stated to be the mode shape vector.  I found somewhere in reading that the residual vector is a scaled vector.  Is this correct?  I figured it couldn't be as there is no input excitation to mass normalize the mode shapes.

5) Is there a realtively convenient and accurate method of transforming the complex mode shapes into normal modes in Labview? Or will I need to use one of the approaches documented in past research and apply it in another program or by hand?

Thank you for any help that you can provide.  I have searched the Discussion forum pervious and had not found any documentation/information on the modal parameter extraction VI's specifically, just how to download and make them compatible with different versions of Labview.

[Paul-B]

Hello bk,

Please take a look at the link below while I look into this further. I will get back with you shortly.

Modal Parameter Extraction LabVIEW VIs

http://zone.ni.com/devzone/cda/epd/p/id/6121

Thank you,

[bk.]

Thank you Paul.  I have read that before but the link doesn't go into any detail about the workings of the SSI VI. 

Thank you in advance for any further information you can provide.

[Paul-B]

Hello bk,

Here is a useful document with information on Stochastic State-Space Modeling

TSA Stochastic State-Space Modeling VI

http://zone.ni.com/reference/en-XX/help/371419D-01/lvtimeseriestk/tsa_stochastic_subspace_modeling/

The default noise subspace is 10%, however if you expect to have increased noise in your signal then the noise subspace percentage should be increased.

In order to transform complex mode shapes to normal mode shapes you will have to manually turn the modes to normal using mathematics.