| Abstract Scope |
Understanding and evaluation of a material's response to an irradiation environment has been a bottleneck in the development, qualification and deployment of new materials and manufacturing technologies in nuclear energy systems. The convergence of high-performance computing, machine learning, high-throughput experimentation, and in situ characterization has the potential to make the materials discovery and diagnosis many times faster than conventional approaches. This talk discusses our recent effort toward machine learning/artificial intelligence (ML/AI)-driven in situ TEM with ion irradiation to rapidly and efficiently assess the irradiation tolerance of a material in nuclear environments. This ML/AI-enabled irradiation diagnosis tool is envisioned to provide real-time microstructure imaging, real-time defect analysis, and real-time property prediction with increasing irradiation dose, and offer a high-fidelity forecast of a material's irradiation tolerance that could dramatically shorten the material discover and development cycles. |