Grainger College of Engineering, All Events

View Full Calendar

Final Doctoral Defense: Majdi Radaideh, PhD Candidate

Event Type
Seminar/Symposium
Sponsor
Department of Nuclear, Plasma, and Radiological Engineering
Location
329 Grainger Engineering Library
Date
Apr 10, 2019   1:00 pm  
Speaker
Majdi Radaideh, PhD Candidate
Cost
Free and Open to the Public
E-Mail
gwitmer2@illinois.edu
Phone
217-333-2295
Views
92
Originating Calendar
NPRE Events

A Novel Framework for Data-driven Modeling, Uncertainty Quantification, and Deep Learning of Nuclear Reactor Simulations

This work presents a modern method for reactor modeling, simulation, and uncertainties through an integrated framework developed under the terminology of combining reactor simulations and experimental data with uncertainty quantification (UQ) and deep learning. The framework houses various physical phenomena that occur inside nuclear reactors, such as neutronics, fuel depletion, thermal-hydraulics, and fuel performance as well as outside the reactor such as spent fuel and criticality safety. The framework utilizes various computer models in the nuclear area and it is supported and validated by a wide range of experimental data in different single and multiphysics experiments, such as delayed neutron data, void fraction measurements, isotopic compositions, nuclear data, and others. The framework is built based upon a wide range of mathematical and statistical methods featuring different areas such as sensitivity analysis, variance decomposition, dimensionality analysis and reduction, reduced order modeling, machine learning, data science, deep learning, uncertainty propagation, Bayesian statistics, correlation analysis, data assimilation, and more. All efforts in this work are expected to yield a better understanding of nuclear reactor simulations, which can lead to improved performance and safety as well as reduced costs for nuclear industry.

The major achievements of the framework developed here to the area include: a set of kinetic parameters' values and uncertainties for reactor systems, advanced models for accurate burnup credit of boiling water reactors, integrated assessment and UQ of nuclear computer models, a platform for nuclear multiphysics simulations, and deep learning models for nuclear high dimensional problems. Most of the methods and frameworks developed in this thesis are extendable to other problems outside the nuclear area.

link for robots only