Benchmark development of temperature dependent critical experiments at the Walthousen reactor critical facility

Eklund, Matthew D.
Thumbnail Image
Other Contributors
Danon, Yaron
Podowski, M.
Lee, Changho
Ji, Wei
Issue Date
Nuclear engineering
Terms of Use
This electronic version is a licensed copy owned by Rensselaer Polytechnic Institute (RPI), Troy, NY. Copyright of original work retained by author.
Full Citation
Critical reactor experiments have been a crucial part of the development of reactor science and technologies in the past 60 years. These experiments serve for the greater understanding of nuclear physics with a reactor and the prediction of reactor operation behavior. More recently, critical experiments have served to validate reactor simulation tools such as MCNP and Serpent 2. There is a continuing need for newer and more accurate critical reactor experiments for the purpose of validation of novel modeling techniques and simulation toolkits; while many critical experiments already exist, many of them have large uncertainties in output criticality, do not have highly accurate system parameters which allow for users to model these experiments, or they don’t provide precise temperature measurements in the system. The Walthousen Reactor Critical Facility (RCF) is a low-power open pool research reactor with the unique capability of performing a wide range of critical experiments. Reactivity experiments of the RCF have been performed and reported using two standard configurations utilizing 332 and 333 fuel pins as well as a non-standard configuration. This non-standard configuration includes a large pipe which was inserted in the center of the reactor core such that heated or cooled water could be pumped inward and outward from the reactor core which served as a coupled physics experiment configuration. However, for this work the heated or cooled water injection experiments are not considered, but the isothermal reactivity experiments which were taken of this configuration are studied. These experiments were performed by carefully raising the reactor tank moderator – light water – using immersed electric heaters, raising the reactor control rods, and measuring the reactivity of the system as a function of temperature. Since the temperatures were slowly raised over the course of multiple hours at a rate of less than 5 degrees Celsius per hour and the moderator was thoroughly mixed using a tank agitator, the thermocouple measurements in the reactor core revealed that the temperature was to within 0.33 degrees Celsius throughout the core. Thus, these experiments are considered to be isothermal. This slow and methodical measurement of reactivity as a function of isothermal temperature is suitable for the development of a set of benchmarks which is the purpose of this work. Contained in this work is the description of the computational models created in Serpent 2 and MCNP and the accompanying sensitivity analysis using these models for the purpose of the development of benchmarks based on the RCF critical experiments. The sensitivity analysis for the standard core configuration revealed there is a ± 166 pcm (per cent mille) benchmark model uncertainty due to input parameters (geometry, material specifications, and temperature measurements) and a ± 191 pcm benchmark model uncertainty for the coupled physics experiment core configuration. Using these computational models, the simulated criticality of the experiments as a function of temperature are compared to experimental measurements and agree to within these benchmark model uncertainties. Since the uncertainty of the criticality as determined in this sensitivity analysis is far less than 1% of the measured values (being within 0.2% of the measured values), and the computational models are within this benchmark uncertainty, this work demonstrates that this dataset is valuable and recommended for the use of validation of reactor simulation tools. Comparing the final benchmark uncertainties of the RCF temperature-dependent critical experiments to other benchmark experiments included in the International Handbook of Evaluated Criticality Safety Benchmark Experiments (IHECSBE) [1] demonstrates the quality of the RCF benchmark dataset developed in this work. The SPERT-D aluminum-clad plate-type fuel experiments (IHECSBE identification numbers HEU-MET-THERM-006 and HEU-MISC-THERM-001) had benchmark uncertainties up to ± 610 pcm, and a set of experiments using SPERT III stainless-steel-clad plate-type fuel in water (IHECSBE ID number HEU-COMP-THERM-022) had a benchmark uncertainty of ± 810 pcm. By comparison, the RCF benchmark uncertainties of ± 166 pcm and ± 191 pcm are exceptional. This demonstrates that the RCF model can be created to a much higher level of accuracy based on the lower uncertainty on criticality due to uncertainty in the model input parameters. The temperature range over which the RCF critical experiments were performed is much wider than the other SPERT experiments mentioned; the SPERT-D aluminum-clad plate-type fuel experiments were measured at 22.2 °C, and the SPERT III stainless-steel-clad plate-type fuel experiments were measured between 55 and 60 °F (approximately between 13 and 16 °C). By comparison, the RCF 333-pin experiments ranged from 29 °C to 46 °C, the 332-pin experiments ranged from 16 °C to 36 °C, and the coupled physics experiment configuration measurements ranged from 11 °C to 41 °C. With such a wide range of temperatures for benchmark measurements, including those below room temperature, this data is highly valuable for validation of computational tool methods for accounting for change in reactor temperature as well as evaluating thermal scattering law/S(α,β) data libraries.
December 2021
School of Engineering
Dept. of Mechanical, Aerospace, and Nuclear Engineering
Rensselaer Polytechnic Institute, Troy, NY
Rensselaer Theses and Dissertations Online Collection
Restricted to current Rensselaer faculty, staff and students in accordance with the Rensselaer Standard license. Access inquiries may be directed to the Rensselaer Libraries.