Reactor Physics

Ph.D. in Nuclear Engineering, UC Berkeley

My doctoral work extended SERPENT 2 (an industry-standard continuous-energy Monte Carlo reactor physics code written in C) by roughly 30,000 lines, producing a first-of-its-kind fuel-cycle depletion capability for circulating, liquid-fueled reactors. Six publications and a granted US patent, plus conference presentations including the International Serpent User Group Meeting.

Since 2020 I've shipped large-scale numerical and machine-learning systems into production under regulatory audit, six years of apprenticeship in being made to prove things. I'd like to bring that back to reactor analysis — a domain where the verification burden was never the afterthought it tends to be elsewhere.

Next

ADER

Advanced Depletion Extension for Reprocessing

Source level rather than input-deck level.
ADER modifies the SERPENT 2 C source directly, a choice that set the standard everything else had to meet: roughly 15,000 lines of unit and integration tests in C and Python, a complete user manual and API, all of it under version control from the first commit. I presented it to the international Serpent community at the 5th Serpent User Group Meeting in Knoxville.
A first-of-its-kind capability
At its core is a chemistry-aware linear optimization engine that operates on the depletion state, allowing fuel composition to be solved against physical and chemical constraints rather than assumed as an input. Only a redacted version of the source is public.
Next

Why I Wrote It

The R in ADER

I wrote ADER because the fuel-cycle tools available for reactors that move and reprocess their fuel rested on assumptions I could not defend, and I wanted depletion results I would be willing to put in front of a regulator.

In a recycling fuel cycle, composition stops being a given and becomes a decision variable coupled to the depletion state — a constrained optimization problem sitting inside a Monte Carlo transport calculation. That is the shape of the problem ADER was built to solve. Any program pursuing fuel recycling alongside the reactor itself is working on that same problem, an endeavor my doctorate was given over to almost entirely.

Next

Applied Nuclear Experience

Oak Ridge National Laboratory · TerraPower · TVA

ORNL, Reactor & Nuclear Systems Division. Researched and documented computational modeling options for kinetic response simulation in circulating-fuel reactors. The resulting survey became a peer-reviewed review paper co-authored with ORNL's Jeffrey J. Powers.

TerraPower, reactor safety. Sodium fast reactor work. I built an automated Python RELAP5 input generator for the safety group, removing hand-construction of thermal-hydraulic input decks.

Tennessee Valley Authority, reactor analysis. Designed a BWR control blade history monitoring and assessment library, with data scrapers running against live reactor operational data.

Next

What Six Years in Production Bought

Numerics, verification, and the habit of having to prove it.

Verification Under Audit

Six years validating stochastic models in a domain where every decision is auditable to a regulator — including an automated evaluation harness for exactly that purpose.

Numerics at Scale

C, C++, Python and Fortran; MPI, OpenMP and CUDA; constrained optimization, sequential Monte Carlo and Bayesian inference on real hardware.

Modern Tooling

A working command of current ML and LLM tooling, and a fairly clear idea of which reactor-analysis problems it is genuinely good for, a shorter list than the enthusiasm around it suggests.

If your program is building something that moves its fuel, I would enjoy talking about where our problems overlap.

Publications & Presentations

  1. Wooten, D., & Fratoni, M. (2020). “Linear optimization for predicting material compositions in molten salt reactors.” Annals of Nuclear Energy, 139, 107159.
  2. Wooten, D., & Powers, J.J. (2018). “A review of molten salt reactor kinetics models.” Nuclear Science and Engineering, 191(3), 203–230. doi:10.1080/00295639.2018.1480182
  3. Shannon, S.C., Knappe, D., Byrns, B., Wooten, D., & Lindsay, A. (2016). U.S. Patent No. 9,475,710. Washington, DC: U.S. Patent and Trademark Office.
  4. Byrns, B., Wooten, D., Lindsay, A., & Shannon, S. (2012). “A VHF driven coaxial atmospheric air plasma: electrical and optical characterization.” Journal of Physics D: Applied Physics, 45(19), 195204.
  5. Byrns, B., Wooten, D., & Shannon, S. (2011). “Design and characterization of a novel coaxial VHF plasma source for air plasma formation.”
  6. Wooten, D., Byrns, B., & Shannon, S. (2011). “RF atmospheric plasma based air filtration using porous metals.” IEEE International Conference on Plasma Science.

Conference Presentations

  • “ADER: Advanced Depletion Extension for Reprocessing — A SERPENT 2 Mod.” 2nd Annual Molten Salt Reactor Workshop, Oak Ridge, TN, 2016.
  • “Enhanced Fidelity Depletion for Molten Salt Reactors.” PHYSOR 2016, Sun Valley, ID.
  • “Advanced Depletion Extension for Reprocessing (ADER) for SERPENT 2.” 5th International Serpent User Group Meeting, Knoxville, TN, 2015.

Dissertation

Wooten, D. (2019). Predicting Fuel Salt Composition via Linear Optimization in Molten Salt Reactors. University of California, Berkeley.

Fellowships & Honors

  • NSSC Fellow — Nuclear Science and Security Consortium, 2019
  • U.S. Nuclear Regulatory Commission Fellow, 2017–2019
  • Berkeley Graduate Fellow, 2016–2017
  • NEUP Fellow — DOE Nuclear Energy University Program, 2013–2016
  • American Nuclear Society Scholarship, 2010–2015

Codes & Methods

SERPENT 2 (code-modification level) · Monte Carlo neutron transport · depletion and fuel cycle analysis · multigroup neutron diffusion · reactor kinetics · RELAP5 · constrained and linear optimization · MPI, OpenMP, CUDA

Contact Me

Email: DanielDavidWooten@gmail.com