Abstract:
This talk will overview a general hybrid framework for reduced order modeling designed to reduce the complexity/dimensionality of multi-scale multi-physics models reactor models in a manner that preserves their accuracy while achieving superior computational performance. Reduced order modeling has been an important subject in many engineering disciplines and has therefore received many advances from different engineering backgrounds. The literature is replete with many techniques that have been proposed for order reduction, e.g., response surface methodologies, proper orthogonal decomposition, variational methods and global sampling methods. Each of these techniques comes with its own advantages and limitations. Our framework has shown that it is possible to combine the benefits of existing techniques using a hybrid approach that can be applied in a generic manner to a general multi-scale multi-physics models. We discuss these developments and demonstrate their applicability to typical reactor design calculations.

Biography:
Dr. Abdel-Khalik is an associate professor at North Carolina State University, Department of Nuclear Engineering. He has obtained his PhD in Nuclear Engineering from NCSU in 2004. Shortly after, he joined AREVA-NP in Lynchburg to work for the PWR's methods group. He returned to Academia and accepted an assistant professor position at NCSU jointly with INL in the fall of 2007. His research interests include inverse problems, uncertainty quantification, reduced order modeling, sensitivity analysis, data assimilation with application to deterministic and probabilistic modeling.