Deciphering STFT

[Dhiraj]

Hi,

Refer to the attached VI (Labview 8). I am making a fiber optic pressure sensor based on photonic crystal fiber for tsunami detection. As the pressure acts on the fiber, the output intensity changes, which I have simulated in the VI. Since this is a signal which varies in both time and frequency, I have used STFT spectrogram for analysis. Unfortunately, I can’t make head or tail of the output!!

Essentially, I wish to find out the wave height (eta) from the output intensity chart.

Can someone please help me?

[johnsold]

I looked at the Help file for the STFT Vi and found an example. The example uses an intensity graph for the output. When I copied that to your VI the output seems to make more sense.

Take a look.

Lynn

[Dhiraj]

Thanks Lynn.
The intensity graph does look better in the sense that it looks less complicated. However, can you help me in correlating the arbit signal eta, that I am currently simulating, with this intensity chart?
What is finally available to me is only the output intensity from which I have to find out the height of the wave (eta in the VI) at the point where the sensor is located.

Incidentally, if you look at the eta that I am generating, at X= 7 and 8, the value is only 0.2 and 0.5 m respectively. This is the height of the tsunami at the said depth (4000m). However, if you look at the corresponding wavelength, (Arbit Signal2 in the VI) at X= 7 and 8, Y = 150 Kms!! This long wavelength is what is distinguishing a tsunami from a surface wave and it is only this tsunami wave which generates a significant dynamic pressure which the sensor is sensing. Even a freak wave with say eta = 20m will not cause any change in the pressure ( only about 10^-37)being recorded at the bottom as its wavelength will be typically only a few meters, say 500m.

Do you think I should employ some kind of a demodulation to the output intensity ( of course I don't have the demodulation toolkit!!)? would be grateful if you could assist me in finding some relation between the STFT output (or any other function) and wave height.
Thanks again.

[johnsold]

I have read about such tsunami detectors. It sounds like an interesting and obviously important project.

I have made some changes to your simulator. These do not necessarily provide an interpretation of the signals, but show how I often approach the analysis of an unknown or poorly known signal.

I eliminated the loop and changed the Arbitrary Signal sources to create the entire waveform at one time (also setting all to the same number of samples). I replaced the Point by Point STFT VI by the standard one. I changed Charts to Graphs. I added a selector to allow easy choice of which signal to analyze.

What I suggest you do next is try different values and see whether you can decide what parameters are most important to your analysis. I do not have enough knowledge of what is meaningful to do that.

Clearly a monitoring program would use the loops and the Point by Point analysis VIs, but I think "batch mode" may be easier for figuring out what you can do.

Lynn