1
|
Liu J, Zhao X, Zhao K, Goncharov VG, Delhommelle J, Lin J, Guo X. A modulated fingerprint assisted machine learning method for retrieving elastic moduli from resonant ultrasound spectroscopy. Sci Rep 2023; 13:5919. [PMID: 37041266 PMCID: PMC10090122 DOI: 10.1038/s41598-023-33046-w] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/03/2023] [Accepted: 04/06/2023] [Indexed: 04/13/2023] Open
Abstract
We used deep-learning-based models to automatically obtain elastic moduli from resonant ultrasound spectroscopy (RUS) spectra, which conventionally require user intervention of published analysis codes. By strategically converting theoretical RUS spectra into their modulated fingerprints and using them as a dataset to train neural network models, we obtained models that successfully predicted both elastic moduli from theoretical test spectra of an isotropic material and from a measured steel RUS spectrum with up to 9.6% missing resonances. We further trained modulated fingerprint-based models to resolve RUS spectra from yttrium-aluminum-garnet (YAG) ceramic samples with three elastic moduli. The resulting models were capable of retrieving all three elastic moduli from spectra with a maximum of 26% missing frequencies. In summary, our modulated fingerprint method is an efficient tool to transform raw spectroscopy data and train neural network models with high accuracy and resistance to spectra distortion.
Collapse
Affiliation(s)
- Juejing Liu
- Department of Chemistry, Washington State University, Pullman, WA, 99164, USA
- Alexandra Navrotsky Institute for Experimental Thermodynamics, Washington State University, Pullman, WA, 99164, USA
- School of Mechanical and Materials Engineering, Washington State University, Pullman, WA, 99164, USA
| | - Xiaodong Zhao
- Department of Chemistry, Washington State University, Pullman, WA, 99164, USA
- Alexandra Navrotsky Institute for Experimental Thermodynamics, Washington State University, Pullman, WA, 99164, USA
| | - Ke Zhao
- Alexandra Navrotsky Institute for Experimental Thermodynamics, Washington State University, Pullman, WA, 99164, USA
| | - Vitaliy G Goncharov
- Department of Chemistry, Washington State University, Pullman, WA, 99164, USA
- Alexandra Navrotsky Institute for Experimental Thermodynamics, Washington State University, Pullman, WA, 99164, USA
- School of Mechanical and Materials Engineering, Washington State University, Pullman, WA, 99164, USA
| | - Jerome Delhommelle
- Department of Chemistry, University of Massachusetts, Lowell, MA, 01854, USA
| | - Jian Lin
- School of Nuclear Science and Technology, Xi'an Jiaotong University, Xi'an, 710049, Shaanxi, China
| | - Xiaofeng Guo
- Department of Chemistry, Washington State University, Pullman, WA, 99164, USA.
- Alexandra Navrotsky Institute for Experimental Thermodynamics, Washington State University, Pullman, WA, 99164, USA.
- School of Mechanical and Materials Engineering, Washington State University, Pullman, WA, 99164, USA.
| |
Collapse
|
2
|
Ma Z, Qi T, Lin L, Lei M. Inverse identification of geometric and acoustic parameters of inhomogeneous coatings through URCAS-based least-squares coupled cross-correlation algorithm. ULTRASONICS 2022; 119:106626. [PMID: 34695748 DOI: 10.1016/j.ultras.2021.106626] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Subscribe] [Scholar Register] [Received: 04/14/2021] [Revised: 10/13/2021] [Accepted: 10/13/2021] [Indexed: 06/13/2023]
Abstract
Aiming to determine the geometric and acoustic parameters of inhomogeneous multi-layer coatings, an ultrasonic pressure reflection coefficient amplitude spectrum (URCAS) is derived using a material-oriented regularization scheme, when wave propagates perpendicularly to the coating surface. Based on the derived theoretical URCAS, a new objective function combining least-squares and cross-correlation algorithm is developed to simultaneously identify the thickness, sound velocity, density, attenuation coefficient amplitude, and attenuation coefficient power-law of the multi-layer coating. Genetic algorithm with the constraint of total multi-layer coating thickness being known is presented to optimize the nonlinear objective function for obtaining global optimal solution. Ultrasonic experiments were implemented on a dual-layer coating specimen with coating1/coating2/substrate structure using a flat immersion transducer with a central frequency of 15 MHz. The inversed thicknesses, sound velocities, and densities of the dual-layer coating were in good agreement with those measured through other methods, with less than 8.1% errors. The inversed attenuation coefficients of the coating 1 and coating 2 were α(f) = 1.02e-2 × f1.93 and α(f) = 4.62e-3 × f1.97, respectively. The upper bounds of the relative errors +r of inverted parameters were all less than 0.061. The proposed ultrasonic inversion method could be used to quantitatively characterize the surface integrity of inhomogeneous multi-layer coatings.
Collapse
Affiliation(s)
- Zhiyuan Ma
- NDT & E Laboratory, Dalian University of Technology, Dalian 116024, China
| | - Tianzhi Qi
- NDT & E Laboratory, Dalian University of Technology, Dalian 116024, China
| | - Li Lin
- NDT & E Laboratory, Dalian University of Technology, Dalian 116024, China.
| | - Mingkai Lei
- Surface Engineering Laboratory, School of Materials Science and Engineering, Dalian University of Technology, Dalian 116024, China
| |
Collapse
|