skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Methodology and Software for Gross Defect Detection of Spent Nuclear Fuel at the Atucha-I Reactor [Novel Methodology and Software for Spent Fuel Gross Defect Detection at the Atucha-I Reactor]

Abstract

Here, fuel assemblies in the spent fuel pool are stored by suspending them in two vertically stacked layers at the Atucha Unit 1 nuclear power plant (Atucha-I). This introduces the unique problem of verifying the presence of fuel in either layer without physically moving the fuel assemblies. Given that the facility uses both natural uranium and slightly enriched uranium at 0.85 wt% 235U and has been in operation since 1974, a wide range of burnups and cooling times can exist in any given pool. A gross defect detection tool, the spent fuel neutron counter (SFNC), has been used at the site to verify the presence of fuel up to burnups of 8000 MWd/t. At higher discharge burnups, the existing signal processing software of the tool was found to fail due to nonlinearity of the source term with burnup.

Authors:
 [1];  [1];  [1];  [2];  [3]
  1. Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
  2. Brazilian-Argentine Agency for Accounting and Control of Nuclear Materials, Rio de Janeiro (Brazil)
  3. National Regulatory Authority - Argentina, Buenos Aires (Argentina)
Publication Date:
Research Org.:
Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1228002
Report Number(s):
LLNL-JRNL-656062
Journal ID: ISSN 0029-5450
Grant/Contract Number:
AC52-07NA27344
Resource Type:
Journal Article: Accepted Manuscript
Journal Name:
Nuclear Technology
Additional Journal Information:
Journal Volume: 192; Journal Issue: 1; Journal ID: ISSN 0029-5450
Publisher:
Taylor & Francis - formerly American Nuclear Society (ANS)
Country of Publication:
United States
Language:
English
Subject:
98 NUCLEAR DISARMAMENT, SAFEGUARDS AND PHYSICAL PROTECTION; 11 NUCLEAR FUEL CYCLE AND RUEL MATERIALS; 21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS; Spent Nuclear Fuel; Gross Defects; Nuclear Safeguards

Citation Formats

Sitaraman, Shivakumar, Ham, Young S., Gharibyan, Narek, Peixoto, Orpet J. M., and Diaz, Gustavo. Methodology and Software for Gross Defect Detection of Spent Nuclear Fuel at the Atucha-I Reactor [Novel Methodology and Software for Spent Fuel Gross Defect Detection at the Atucha-I Reactor]. United States: N. p., 2017. Web. doi:10.13182/NT14-63.
Sitaraman, Shivakumar, Ham, Young S., Gharibyan, Narek, Peixoto, Orpet J. M., & Diaz, Gustavo. Methodology and Software for Gross Defect Detection of Spent Nuclear Fuel at the Atucha-I Reactor [Novel Methodology and Software for Spent Fuel Gross Defect Detection at the Atucha-I Reactor]. United States. doi:10.13182/NT14-63.
Sitaraman, Shivakumar, Ham, Young S., Gharibyan, Narek, Peixoto, Orpet J. M., and Diaz, Gustavo. Mon . "Methodology and Software for Gross Defect Detection of Spent Nuclear Fuel at the Atucha-I Reactor [Novel Methodology and Software for Spent Fuel Gross Defect Detection at the Atucha-I Reactor]". United States. doi:10.13182/NT14-63. https://www.osti.gov/servlets/purl/1228002.
@article{osti_1228002,
title = {Methodology and Software for Gross Defect Detection of Spent Nuclear Fuel at the Atucha-I Reactor [Novel Methodology and Software for Spent Fuel Gross Defect Detection at the Atucha-I Reactor]},
author = {Sitaraman, Shivakumar and Ham, Young S. and Gharibyan, Narek and Peixoto, Orpet J. M. and Diaz, Gustavo},
abstractNote = {Here, fuel assemblies in the spent fuel pool are stored by suspending them in two vertically stacked layers at the Atucha Unit 1 nuclear power plant (Atucha-I). This introduces the unique problem of verifying the presence of fuel in either layer without physically moving the fuel assemblies. Given that the facility uses both natural uranium and slightly enriched uranium at 0.85 wt% 235U and has been in operation since 1974, a wide range of burnups and cooling times can exist in any given pool. A gross defect detection tool, the spent fuel neutron counter (SFNC), has been used at the site to verify the presence of fuel up to burnups of 8000 MWd/t. At higher discharge burnups, the existing signal processing software of the tool was found to fail due to nonlinearity of the source term with burnup.},
doi = {10.13182/NT14-63},
journal = {Nuclear Technology},
number = 1,
volume = 192,
place = {United States},
year = {Mon Mar 27 00:00:00 EDT 2017},
month = {Mon Mar 27 00:00:00 EDT 2017}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Save / Share:
  • At the Atucha-I pressurized heavy water reactor in Argentina, fuel assemblies in the spent fuel pools are stored by suspending them in two vertically stacked layers. This introduces the unique problem of verifying the presence of fuel in either layer without physically moving the fuel assemblies. Since much of the fuel is very old, Cerenkov viewing devices are often not very useful even for the top layer. Given that the facility uses both natural uranium and slightly enriched uranium at 0.85 w% {sup 235}U, and has been in operation since 1974, a wide range of burnups and cooling times canmore » exist in any given pool. A spent fuel neutron counting tool consisting of a fission chamber, SFNC, has been used at the site to verify the presence of fuel up to burnups of 8000 MWd/t. At higher discharge burnups to levels up 11,000 MWd/t, the existing signal processing software of the tool was found to fail due to non-linearity of the source term with burnup. A new Graphical User Interface software package based on the LabVIEW platform was developed to predict expected neutron signals covering all ranges of burnups and cooling times and establish maps of expected signals at various pool locations. The algorithm employed in the software uses a set of transfer functions in a 47-energy group structure which are coupled with a 47-energy group neutron source spectrum based on various cooling times and burnups for each of the two enrichment levels. The database of the software consists of these transfer functions for the three different inter-assembly pitches that the fuel is stored in at the site. The transfer functions were developed for a 6 by 6 matrix of fuel assemblies with the detector placed at the center surrounded by four near neighbors, eight next nearest neighbors and so on for the 36 assemblies. These calculations were performed using Monte Carlo radiation transport methods. The basic methodology consisted of starting sources in each of the assemblies and tallying the contribution to the detector by a single neutron in each of the 47 energy groups used. Thus for the single existing symmetric pitch in the pools, where the vertical and horizontal separations are equal, only 6 sets of transfer functions are required. For the two asymmetrical pitches, nine sets of transfer functions are stored. In addition, source spectra at burnups ranging from 4000 to 20000 MWd/t and cooling times up to 40 years are stored. These source terms were established based on CANDU 37-rod fuel that is very similar to the Atucha fuel. Linear interpolation is used by the software for both burnup and cooling time to establish source terms at any intermediate condition. Using the burnup, cooling time and initial enrichment of the surrounding assemblies a set of source strengths in the 47-group structure for each of the 36 assemblies is established and multiplied group-wise with the appropriate transfer function set. The grand total over the 47 groups for all 36 assemblies is the predicted signal at the detector. The software was initially calibrated against a set of typically 5-6 measurements chosen from among the measured data at each level of the six pools and calibration factors were established. The set used for calibration is chosen such that it is fairly representative of the range of spent fuel assembly characteristics present in each level. Once established, these calibration factors can be repeatedly used for verification purposes. Recalibration will be required if the hardware or pool configurations has changed. It will also be required if a long enough time has elapsed since they were established thus making a cooling time correction necessary. The objective of the inspection is to detect missing fuel from one or more nearest neighbors of the detector. During the verification mode of the software, the predicted and measured signals are compared and the inspector is alerted if the difference between the two signals is beyond a set tolerance limit. Based on the uncertainties associated with both the calculations and measurements, a lower limit of the tolerance will be 15% with an upper limit of 20%. For the most part a 20% tolerance limit will be able to detect a missing assembly since in the vast majority of cases the drop in signal due to a single missing nearest neighbor assembly will be in the range 24-27%. The software was benchmarked against an extensive set of measured data taken at the site in 2004. Overall, 326 data points were examined and the prediction of the calibrated software was compared to the measurements within a set tolerance of ±20%. Of these, 283 of the predicted signals representing 87% of the total matched the measured data within ±10%. A further 27 or 8% were in the range of ±10-15% and 8 or 2.5% were in the range of ±15-20%. Thus, 97.5% of the data matched the measurements within the set tolerance limit of 20%, with 95% matching measured data with the lowest allowed tolerance limit of ±15%. The remaining 2.5% had measured signals that were very different from those at locations with very similar surrounding assemblies and the cause of these discrepancies could not be ascertained from the measurement logs. In summary, 97.5% of the predictions matched the measurements within the set 20% tolerance limit providing proof of the robustness of the software. This software package linked to SFNC will be deployed at the site and will enhance the capability of gross defect verification for the whole range of burnup, cooling time and initial enrichments of the spent fuel being discharged into the various pools at the Atucha-I reactor site.« less
  • At the Atucha-I pressurized heavy water reactor in Argentina, fuel assemblies in the spent fuel pools are stored by suspending them in two vertically stacked layers. This introduces the unique problem of verifying the presence of fuel in either layer without physically moving the fuel assemblies. Movement of fuel, especially from the lower layer, would involve a major effort on the part of the operator. Given that the facility uses both natural uranium and slightly enriched uranium at 0.85 w% {sup 235}U, and has been in operation since 1974, a wide range of burnups and cooling times can exist inmore » any given pool. Additionally, while fuel assemblies are grouped together in a uniform fashion, the packing density from group to group can vary within a single pool. A tool called the Spent Fuel Neutron Counter (SFNC) was developed and successfully tested at the site to verify, in an in-situ condition, the presence of fuel up to burnups of 8,000 MWd/t. Since the neutron source term becomes a nonlinear function of burnup beyond this burnup, a new algorithm was developed to predict expected response from the SFNC at measurement locations covering the entire range of burnups, cooling times, and initial enrichments. With the aid of a static database of parameters including intrinsic sources and energy group-wise detector response functions, as well as explicit spent fuel information including burnups, cooling times, enrichment types, and spacing between fuel assemblies, an expected response for any given location can be calculated by summing the contributions from the relevant neighboring fuel assemblies. Thus, the new algorithm maps the expected responses across the various pools providing inspectors with a visual aid in verifying the presence of the spent fuel assemblies. This algorithm has been fully integrated into a standalone application built in LabVIEW. The GUI uses a step-by-step approach to allow the end-user to first calibrate the predicted database against a set of measurements with SFNC at selected locations where spent fuel is present. Once the database is calibrated it can be used to detect gross defects by comparing the measured signal to the one predicted by the database with differences beyond a set tolerance indicating missing fuel.« less
  • In this paper we seek to create a model by determining the field of view (FOV) of a detector (i.e. which assemblies contribute to the detector response) in the Atucha-I spent fuel pool. The FOV is determined by solving the adjoint transport equation using the 3-D, parallel PENTRAN (Parallel Environment Neutral-particle TRANsport) Sn code, with the detector cross section as the adjoint source. If this adjoint function is coupled with the source spectrum, then the contribution to the detector from each assembly can be determined. First, the reactor criticality was modeled using the MCNP5 (Monte Carlo N-Particle) Monte Carlo codemore » in order to determine the power distribution in each assembly. Using the power distribution data, the assemblies were divided and homogenized into 8 axial and 3 radial zones for burnup analysis. Depletion calculations were performed for each zone using the ORIGEN-ARP (Automatic Rapid Processing) utility from the SCALE 5.1 (Standardized Computer Analyses for Licensing Evaluation) code package. Spent fuel pool and detector were modeled in 2-D in PENTRAN as the detector plus 3 fuel assemblies along both x and y axes. Using the resulting adjoint function combined with the source spectrum, they have determined the FOVs of the fission chamber neutron detector that was used at Atucha, and concluded that 2 assemblies along x and y axes are needed for both cases (i.e. the 4 adjacent assemblies plus the next surrounding 12). For the neutron detector, 88% of the response comes from the nearest 4 assemblies, with 99% from the nearest 16. Results for a uniformly sensitive gamma detector indicate that 2 assemblies in both directions are also needed, with 89% of the response coming from the adjacent assemblies. A Monte Carlo calculation using MCNP was performed to benchmark the neutron result, giving a similar result (87% MCNP vs. 88% PENTRAN). Based on these studies, we have developed a database of FOVs as a function of burnup and decay conditions for different detector types, and a methodology/algorithm which uses this database to analyze the response of a detector placed in a spent fuel pool with the aim of detecting gross defects.« less
  • An improved system was developed to recover the initial parameters and operating conditions from spent nuclear fuel. This work is an expansion of a previous proof-of-concept system developed by the author. The improved system increases the fidelity of the forward model within the spent fuel forensic inverse analysis using two unique methodologies. The first improvement consists of developing a system to accurately create one-group neutron crosssection libraries for any user-specified reactor system. As such, a detailed model using the depletion code MONTEBURNS is developed. During MONTEBURNS execution, cross-section libraries are generated at every user specified burnup step in time. Thesemore » libraries could be developed for many reactor systems, then housed in a database and used for analyzing spent fuel.« less