Modal Identification SSI method

[Rafo]

Dear all,

I am working in labview 8.0. I am developing a tool to evaluate the dynamic response of civil structures using the stochastic subspace identification method based in output only tests (I acquire accelerations due by ambient loads).

I have two problems:
a) The first one is that when I run the labview example "Modal Analysis of a plate.vi" the results of damping factors are negative! as you can know it is impossible.
b) The second one is that I adapted this example to my requirements but the results of frequency estimation vary in order of 10% in each run of the same point which as you can imagine is too much (The damping factors are negative again). I think I am doing something wrong could anybody help me? I am attaching the vi.

Thank you very much

Rafo

[Mathieu_R.]

Hi Rafo,

If you think the example doesn't give coherent results, why do you are trying to adapt it for your application? I'm not a specialist of this kind of analysis, but you should start on a base you trust.

About 10% variation, this may be explain acquisition scale. If you acquire very small signal with +-10v scale, errors of this kind may appear. What are the settings of your task?

Cordially,

[Rafo]

Thank you ... my task is configured like this:

Sensitivity: 10000 mV/g
Iex value (A) : 2 m
db reference: 1
Sampling rate :2 k
Samples to read : 1 k
Signal imput range : +-5 mg

[Rafo]

Dear Mathieu,

I am trying to use the TSA Stochastic State-Space Modelling tool. I am just using this tool without any modification. I think that in order to get good results of frequencies it is necessary signals with high amplitudes in the frequencies of interests. The problem is that the damping factors are again very bad and still negatives which means that these values are not good. Do you know if I need to pre-process the data with some tool before applying the tool for modal identification?

Thanks, 

Rafo

[TinoMK]

Hi Rafo,

Have you solved your problem? I am trying to do the same thing. 

Thanks,