This project’s scientific research goal is to develop simulation and modeling foundations for pulsed laser (PL) induced guided ultrasonic waves (GUW) and their propagation in nuclear energy relevant structural components, and to develop machine learning (ML) based computational data interpretation methods to support the understanding of the wave data. The project has four research objectives toward achieving its goal: developing a physics-based temporal-spatial thermos-elastic model for pulsed laser induced GUW excitation, developing an efficient hybrid modeling of global wave propagation and local wave-discontinuity interactions, performing experimental validation of the model, and creating a multi-domain ML framework to assist robust and effective evaluation and characterization of wave propagation and structural discontinuity.

For the PL actuation model, we extend traditional optical-thermal-elastic transduction model by including temporal and spatial distribution functions for laser optical input and structural material properties in the thermal conductivity equation to study how laser characteristics and material properties will affect the resulted GUW modes and frequency contents. The salient hybrid modeling of wave propagation is achieved by judiciously combining the efficiency of analytical modeling of global wave propagation and the high accuracy of wave-discontinuity interactions with finite element modeling (FEM). The laboratory model validation is conducted with state of the arts pulsed laser device as well as scanning laser Doppler vibrometer to ensure the simulated waves contain critical wave features such as dispersion behaviors and frequency contents. The multi-domain Explainable ML (xML) framework is based on Transformer and Memory Bank to assist robust and effective understanding of laser GUW signals, and hence improves detection and characterization of propagation and discontinuity in structural components. The technical outcomes of this project will be an efficient and effective simulation suite to enable in-depth understanding of wideband and high frequency PL excited GUW and their propagations in structures, driven by computationally powerful ML based data analysis. The work will not only establish the foundation for adopting PL GUW for structural inspection, but also provide an efficient modeling and predictive framework for in-depth study of GUW excitation and propagation mechanisms and develop robust wave data analysis algorithms based on the cutting-edge ML technology.

Within this project, our goal of promoting inclusive and equitable research (PIER) is developing and implementing a far-reaching PIER model that builds upon a multidimensional partnership and three pillars of intentional efforts that are designed to strengthen diverse research excellence. Our first PIER objective is to build up the multidimensional partnership that includes (a) an academic partnership between an established research university and a four-year Historically Black Colleges and Universities, (b) an academic-national laboratory partnership, and (c) a diverse research team and committed federal agency partnership. With the structured partnership, our second PIER objective is creating sustainable diverse research excellence through the three pillars of intentional activities that focus on creating a clear pathway for researchers with underrepresented backgrounds to begin pursuing scientific research, to excel in their research pursuits, and achieve a persistent dedication to energy related research and careers. Through our PIER efforts, we will demonstrate a model of collective and synergic efforts that aims to maximize resources, augment outcomes and hence accelerate PIER progress, supporting the creation of diverse pathways to excellence and benefiting both the DOE research communities and the society.