Benchmark data set validates global nuclear reactor codes
Published in the Annals of Nuclear Energy, the re-analysis used improved radiochemical methods developed at ORNL by Joe Giaquinto with Jeffrey Delashmitt and Tamara Haverlock of ORNL’s Chemical Sciences Division and characterized more than 50 different isotopes and 16 elements in the fuel rods with high accuracy. It produced an experimental data set with uncertainties many times smaller than those obtained by the earlier radiochemical analysis. With this improved standard of experimental accuracy, modeling and simulation experts at ORNL — first author Ian Gauld, Jianwei Hu and Germina Ilas, all of the Reactor and Nuclear Systems Division — applied the radiochemical analysis and NDA data to validate codes widely used by the nuclear safeguards research community.
Improving experimental accuracy
“We created a benchmark for other nuclear analytical chemistry labs to follow for detailed used fuel characterizations,” said Giaquinto, who led the chemical analytics of the new study. “We can provide experimental data with a much improved degree of confidence that modelers can use to evaluate predicted results from their computational models. This gives them a more accurate data set when validating reactor codes based on the empirical data. We improved the experimental end of the validation process.”
Giaquinto and colleagues procured the key lanthanide standards that made the high-precision measurements possible from the National Isotope Development Center (NIDC), whose Business Office is located at ORNL. The NIDC, which is supported by the DOE Isotope Program (managed by the Office of Science), coordinates isotope production and sales across DOE Isotope Program–funded facilities. Giaquinto’s team purchased additional standards from a commercial vendor.
Combining high-pressure, high-resolution separation techniques with isotopically enriched standards, Giaquinto, Delashmitt and Haverlock were able to rapidly isolate multiple elements in sets of actinides and lanthanides. The elementally pure fractions were then analyzed using isotope dilution, the most accurate technique available for quantification with mass spectrometry.
“For isotope dilution, a well certified enriched isotope of an element is added gravimetrically to the working solution, or spiked, before chemical processing,” Giaquinto explained. “The enriched isotope is now the calibrant and is equilibrated into the system with the unknowns. From there on out, whatever you do to that sample, the calibrant follows the same chemistry process as the unknown, allowing for a high degree of precision in the measurement compared to an external calibrant.”
That technique produced uncertainties of 1.5 to 2 percent, a dramatic improvement over uncertainties of 10 percent or more produced with previous techniques.
The scope of nuclear resources at ORNL made the achievement possible. “We have a facility where they could handle the fuel,” Giaquinto said. “We have hot cells where we could do detailed dissolution. And we have the breadth of nuclear expertise from reactor modeling to NDA to analytical chemistry.”
Capturing complexities using modeling and simulation codes
Simulation codes are used to model the complexities of a reactor and track the creation and destruction of more than 1,600 nuclides in nuclear fuel during its life in the reactor. The reactor core contains more than a hundred fuel assemblies, each in turn filled with hundreds of zirconium-alloy-clad uranium oxide rods that heat water to generate power. Each fuel rod experiences a different history in the reactor resulting from its unique position in the fuel assembly.
Researchers widely use the SCALE nuclear systems modeling and simulation code package, developed at ORNL, to model fuel assembly designs and complex reactor operating conditions and calculate changes in the isotopic content of nuclear fuel as it is burned. Using ORNL’s new measurement data for the TMI-1 fuel samples, the current study yielded excellent agreement between code predictions and measurements and provided more accurate inventory data for comparison with NDA safeguards measurements. Moreover, the new high-precision measurements have resolved discrepancies between previous measurements and simulated values for some isotopes in one of the TMI-1 cornerstone data sets that have perplexed the scientific community for more than 10 years.
Gauld’s code-validation work is part of a global effort to compare simulation results against data from experiments conducted internationally. Representing DOE, Gauld chairs the Organization for Economic Co-operation and Development’s Nuclear Energy Agency subgroup on assay data for used nuclear fuel, which is compiling a public database of measurements taken on more than 600 fuel samples. The information from the ORNL study will be entered into that database. “We want to capture all measurements of this type that have been done in the world for different reactor types and make it a central, go-to database for experiment used fuel inventory data,” Gauld said. The database will provide not only measurement data, but also fuel design information and all primary references needed for modeling. It will reside both at the NEA Data Bank in Paris and at the Radiation Safety Information Computational Center in Oak Ridge.
DOE’s Office of Civilian Radioactive Waste Management funded initial measurements of the TMI-1 fuel rods made between 1998 and 2000 and ORNL’s acquisition of several of the TMI-1 fuel rod segments in 2008 for destructive assay measurement. Subsequent funding from DOE’s National Nuclear Security Administration, Office of Defense Nuclear Nonproliferation Research and Development, supported NDA and analytical chemistry measurements of the fuel, and code-validation efforts in support of improved nuclear spent fuel safeguards.
— Read more in I. C. Gauld et al., “Re-evaluation of spent nuclear fuel assay data for the Three Mile Island unit 1 reactor and application to code validation,” Annals of Nuclear Energy 87, no. 2 (January 2016): 267-81 (DOI: 10.1016/j.anucene.2015.08.026)
