Tanja Kovacevic

MIT Department: Nuclear Science and Engineering
Undergraduate Institution: University of Colorado Denver
Faculty Mentor: Emilio Baglietto
Research Supervisor: Micheal Acton
Website: LinkedIn

2019 Research Poster


My name is Tanja Kovacevic and I am a rising senior at the University of Colorado at Denver. I am Chemistry major with a minor in Mathematics. My goal is to pursue a PhD in the nuclear/chemical sciences and computationally model systems which will contribute to the application of nuclear fission as a source of clean energy. I hope to help mitigate the energy crisis and move the world towards the use of cleaner energy sources. Outside of school I enjoy working out, going to events around town, and enjoying free time with friends and family.

2019 Research Abstract

Assessment of a Physics Based UQ Method for the Application of CFD

 Tanja Kovacevic1, Michael Acton2, and Emilio Baglietto2
1Department of Chemistry, University of Colorado Denver
2Department of Nuclear Science & Engineering, Massachusetts Institute of Technology

Nuclear as a source of clean energy supports climate change efforts and is a ready alternative to fossil fuel. Unfortunately, low cost gas hinders the competitiveness of nuclear power. Improving economics, through better operations is the key to support the existing nuclear industry framework. Computational modeling is a viable method in modeling the complex behaviors of fluid within reactors without the need for expensive experimentation. Predicting behavior of turbulence in nuclear reactors has proven a significant challenge. Computational fluid dynamics (CFD) is a powerful tool used to model 3D turbulent flow in a reactor and allows researchers to investigate fluid behaviors not available to the naked eye. Yet, any experiment or simulation performed quantitatively must take into account uncertainty to ensure quality of measurements. Uncertainty sources ranging from boundary conditions to physical models lead to significant error. The ability to represent uncertainty is imperative in moving the nuclear industry forward. Our initial studies determine uncertainty quantification (UQ) of a numerical benchmark, GEMIX, on three experimental profiles: mean concentration, turbulent kinetic energy (TKE), and velocity. The goal of the developed UQ method is to capture experimental results within the uncertainty bounds. The ability to bound experimental data and uncertainty makes CFD simulations a robust computational model with the ability to produce accurate results without experimentation. Using CFD alongside the proposed UQ method, simulations now offer an economic feasibility in sourcing clean energy from nuclear power.