Session: 01-10-01: Congress-Wide Symposium on NDE & SHM: Ultrasonic Waves for Material Characterization and Damage Assessment
Paper Number: 90232
90232 - 2D Numerical Ultrasound Computed Tomography for Elastic Material Properties in Metals
Adequate knowledge about a material through characterization during the development, production, and processing of the material is required for quality assurance and in-service safety. Material characterization involves the evaluation of properties such as elastic coefficients, material microstructures, morphological features, and associated mechanical properties. Ultrasonic signals are sensitive to useful acoustic properties, including wave speeds, attenuation, diffusion backscattering, variations in microstructure, and elastic properties (e.g., elastic modulus, hardness, etc.). To obtain a quantitative estimation of the material properties, an emerging imaging technique known as ultrasound computed tomography (USCT) can be utilized. This paper proposes to map the wave speeds (i.e., longitudinal and shear) inside elastic parts employing a wave-based methodology (known as full waveform inversion (FWI)) for USCT.
FWI is a partial differential equation-constraint, nonlinear optimization technique. It is based on full wave-field modeling and inversion to extract material parameter distribution using wave equations. FWI can consequently produce high-resolution images by iteratively determining and minimizing a waveform residual, which is the difference between the modeled data and the acquired data.
A multi-channel ultrasound data acquisition system was used to acquire data using full matrix capture (FMC) in transmission mode from an aluminum sample, where two 500 kHz linear phased array transducers served as sources and receivers respectively. The sources and receivers were aligned at the top and bottom of the specimen, respectively. The acquired signals were later being preprocessed and used for the reconstruction of the final model. The inversion was performed using a Python-based workflow.
In this study, we initially developed a synthetic model using a 2D spectral finite element-based solver (i.e., SPECFEM2D) for preliminary benchmarking. A true model was created where the material properties (i.e., longitudinal, and shear wave speed distribution) were considered to be unknown, and accordingly, an initial model with known material properties was assumed. We analyzed the performance of FWI of the benchmarked model. The accurate forward modeling of the acquired experimental signals is a pivotal part to generate high-resolution FWI images. As the actual source information is not available to us due to the unknown transducer transfer functions, we need to perform source inversion to estimate the accurate source information from the acquired signals which strongly impact the quality of the reconstructed images. While inverting for the source, a linear inversion formula was implemented utilizing acquired and synthetic data. Several critical processes contribute to the success of this approach: selection of the synthetic and observed time window, selection of the source frequency in the synthetic model. The time delay of the inverted source signal was adjusted by calculating the travel time difference between the synthetic and acquired signal.
Finally, the inverted source time function and the acquired signals were fit into the proposed FWI model to perform the reconstruction. Reconstructed images were analyzed with respect to the actual elastic specimen. The performance of FWI based ultrasound tomography in material characterization that has been presented in this study exhibits application potentials in nondestructive material characterization.
Presenting Author: Jiaze He University of Alabama
Presenting Author Biography: Dr. Jiaze He is an assistant professor in Aerospace Engineering and Mechanics(AEM)at the University of Alabama. His lab develops/implements novel ultrasonic imaging methods for materials/structures/medical imaging. Before he arrived at the UA, he was a postdoctoral research associate in the Theoretical & Computational Seismology Group at Princeton University and a research scholar at NASA LaRC in the Digital Twin program. Currently, he serves as a member of SAE AMSK Non-Destructive Methods and Processes Committee, Nondestructive Evaluation, Diagnosis, and Prognosis Division (NDPD) EC at ASME, Additive Manufacturing for Maintenance Operations Working Group, and AM Common Data Model Group.
Authors:
Md Aktharuzzaman The University of AlabamaShoaib Anwar The University of Alabama
Dmitry Borisov The University of Kansas
Jing Rao University of New South Wales
Jiaze He University of Alabama
2D Numerical Ultrasound Computed Tomography for Elastic Material Properties in Metals
Paper Type
Technical Paper Publication