Hi Amro,

A few points:

1. Change your acquisition settings, you are wasting a lot of data storage space on noise. Also, it makes it difficult to sort through the data to find your actual waveform.

2. On slide 1 of the attached pdf is the front portion of the waveform that you captured. I have identified the start of both of the plate modes for you in the time domain. Attempting to identify the dispersion is a bit more complicated and it appears to be due to the sensor that you are using. I didn't see a mention of sensor type, but judging by the waveform it is somewhat flat, but still has a resonance that is masking the mode shapes. I would venture a guess that you are using a WD type sensor (or something similar), but that is just a guess. For a cleaner analysis I would suggest a flatter sensor (B1025, conical element transducer, etc.)

3. To show the benefits of a flat sensor and a superior time frequency distribution (TFD), slide 2 shows you the time domain waveform and the Smoothed Psuedo Wigner Ville Distribution of the same waveform for a similar experiment. Source was an in-plane PLB on a 3 mm 5132 Al plate, sensor was a Digital Wave B1025, I don't recall propagation distance, but 150 mm seems about right based upon S0 and A0 difference in front end arrival times. From the flat sensor, the mode shapes are easily identifiable in the time domain. From the superior TFD, the energy from the mode shapes don't get spread across all frequencies as they do with the wavelet transform, which makes mode identification difficult. If you are going to use the Vallen software, I would suggest switching to the Choi Williams TFD.

4. I would not try to run your time domain waveform through the software until you cut down the waveform window, 4096 or 8192 points (max) should be more than appropriate for your set-up. You may also consider increasing your sample rate from 2.5 MHz to 5 or 10 MHz to capture the S0 mode reversal on a 1.5 mm steel plate (f ~ 1600 kHz)... Nyquist frequency is stopping you at 1250 kHz currently.

I hope these suggestions are helpful.

Brian

Hi Brian,

Thanks for your helpful reply.

In the first slide you in your pdf. file, I would like to know what makes you judged that S0 and A0 (TOA) is that pointed time? It is clear for me the edge reflections point.

And if I need to calculate the velocity of the S0 and A0 mods what is the frequency I use?

I use Kistler 8152B111 sensor and NI DAQ PCI-6132 (attached the data sheet).

Do you think is that sensor and DAQ suitable for AL plate 6.3mm thickness, as I face the same problem? What is the effect of plate thickness?

I attached the signal data file after I cut down the wave form window, does that what you mean?

Thanks for helping

Amro

Amro,

No problem. I have looked at a fair bit of MAE/Guided Wave signals, picking out the modes becomes easier the more you do it. The front end (low frequencies) of the S0 mode are the fastest propagating, so you would expect them to show up first. In some materials, the higher order modes can have comparable velocities, but they are of limited applicability in most "thin" plate structures (