| Title |
SNR equalization in non-contact resonant ultrasound spectroscopy measurements |
| Authors |
Tayyib, Muhammad ; Svilainis, Linas |
| DOI |
10.1016/j.ndteint.2025.103386 |
| Full Text |
|
| Is Part of |
NDT & E international.. London : Elsevier. 2025, vol. 154, art. no. 103386, p. 1-11.. ISSN 0963-8695. eISSN 1879-1174 |
| Keywords [eng] |
Non-contact resonant ultrasound spectroscopy ; Signal to noise ratio ; Spectral losses compensation ; Spread spectrum signals |
| Abstract [eng] |
It is proposed to replace the single pulse excitation with spread spectrum signal of programmable spectral shape in non-contact resonant ultrasound spectroscopy (NC-RUS). Bandwidth can be improved by pushing the signal energy towards high loss frequencies. It has been demonstrated that not just signal spectrum but also noise spectral density are irregular, therefore not only signal spectrum but SNR has to be equalized. The benefit of SNR-equalized signals in RUS is the improved SNR over broader bandwidth which results in better inverse solution convergence. Arbitrary position, and width pulses (APWP) sequences performed best. It was demonstrated that unipolar APWP signals can be used to obtain the acceptable performance of spectral shaping. Use of unipolar excitation allows for simpler electronics. While −20 dB bandwidth in case of pulse excitation was 480 kHz, APWP excitation produced 670 kHz bandwidth. The proposed was compared against pulse, linear frequency modulation (LFM) and nonlinear frequency modulation (NLFM) signals in NC-RUS measurements of plant leaf and polycarbonate plate. In leaf measurements, bias error in density estimation was 2 % with APWP signals while LFM resulted in 45 % and pulse in 60 % error. The polycarbonate sample velocity estimation error for polycarbonate sample using APWP signal was 0.2 %, while LFM was 10 times higher (2 %); density estimation error for APWP was 3 %, but LFM resulted in 36 %, NLFM error was 22 %, pulse excitation resulted in 12 % error. |
| Published |
London : Elsevier |
| Type |
Journal article |
| Language |
English |
| Publication date |
2025 |
| CC license |
|
