TY - JOUR
T1 - Grain size evaluation with time-frequency ultrasonic backscatter
AU - Bu, Yangguang
AU - Liu, Xiling
AU - Turner, Joseph A.
AU - Song, Yongfeng
AU - Li, Xiongbing
N1 - Funding Information:
This work was supported by the National Key R&D Program of China [grant numbers 2017YFB1201302-13], and the Fundamental Research Funds for the Central Universities of Central South University [grant numbers 2020zzts557].
Funding Information:
This work was supported by the National Key R&D Program of China [grant numbers 2017YFB1201302-13 ], and the Fundamental Research Funds for the Central Universities of Central South University [grant numbers 2020zzts557 ].
Publisher Copyright:
© 2020
PY - 2021/1
Y1 - 2021/1
N2 - The ultrasonic backscatter signal can be used to evaluate the grain size of polycrystalline media nondestructively. In this article, the joint time-frequency analysis is introduced into the ultrasonic backscatter model, with the independent scattering approximation. The time-frequency domain model is applied to predict the spatial variance surface (SVS) and fit the experimental results to evaluate the grain size with the approach giving a relative error of 2.26% in comparison with metallography. Moreover, it shows that the present method suppresses electrical noise better than either the time domain or the frequency domain models alone. Based on the global sensitivity analysis of Sobol's method, the sensitivities of transducer parameters and material parameters in the ultrasonic backscatter model are estimated. The sensitivity analysis shows that not all the transducer parameters in the model must be calibrated carefully. In addition, it is found that the grain size sensitivity index of the time-frequency backscatter model is higher than the time domain or the frequency-domain backscatter models. Overall, the backscatter model in the time-frequency domain can evaluate the microstructure more accurately.
AB - The ultrasonic backscatter signal can be used to evaluate the grain size of polycrystalline media nondestructively. In this article, the joint time-frequency analysis is introduced into the ultrasonic backscatter model, with the independent scattering approximation. The time-frequency domain model is applied to predict the spatial variance surface (SVS) and fit the experimental results to evaluate the grain size with the approach giving a relative error of 2.26% in comparison with metallography. Moreover, it shows that the present method suppresses electrical noise better than either the time domain or the frequency domain models alone. Based on the global sensitivity analysis of Sobol's method, the sensitivities of transducer parameters and material parameters in the ultrasonic backscatter model are estimated. The sensitivity analysis shows that not all the transducer parameters in the model must be calibrated carefully. In addition, it is found that the grain size sensitivity index of the time-frequency backscatter model is higher than the time domain or the frequency-domain backscatter models. Overall, the backscatter model in the time-frequency domain can evaluate the microstructure more accurately.
KW - Grain size evaluation
KW - Sensitivity analysis
KW - Short-time fourier transform
KW - Time-frequency analysis
KW - Ultrasonic backscatter
UR - http://www.scopus.com/inward/record.url?scp=85092451837&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85092451837&partnerID=8YFLogxK
U2 - 10.1016/j.ndteint.2020.102369
DO - 10.1016/j.ndteint.2020.102369
M3 - Article
AN - SCOPUS:85092451837
SN - 0963-8695
VL - 117
JO - NDT and E International
JF - NDT and E International
M1 - 102369
ER -