http://www.diva-portal.org
Gamma Emission Tomography
of LOCA transient test rods
Project title: Development and demonstration of
quantitative detail reconstruction for Gamma Emission Tomography of LOCA test rods.
Swedish title: Utveckling och demonstration av kvantitativ
detaljrekonstruktion för gammaemissionstomografi av LOCA-teststavar
Project leader: Peter Andersson, Uppsala University Contact person on SSM: Anna Alvestav
Duration: 2017-2017
Background
• At the Halden Reactor, there is an ongoing test series for Loss-Of-Coolant-Accident (LOCA) tests with high-burnup nuclear fuel.
• In this Project, we have developed the technique of Gamma Emission Tomography (GET) for quantitative reconstructions of relocated fuel in LOCA test rods.
Gamma Emission Tomography
• The Halden Tomography system uses a single spectroscopic detector to scan the fuel laterally and rotationally • A mathematical reconstruction is
performed of the emission intensity of selected radionuclides.
Andersson, Peter; Holcombe, Scott; ",A computerized method (UPPREC) for quantitative analysis of irradiated nuclear fuel assemblies with gamma emission tomography at the Halden reactor, submitted to Annals of Nuclear Energy, 2017,
1) During LOCA transient fuel and cladding temperatures rise.
2) The cladding expands to a balloon formation in the most affected location.
3) Fuel may relocate to the free volume of the balloon, adding additional local heat load.
4) In high-burnup fuel, fragmentation has been
observed to cause higher degree of relocation and higher packing fraction in the balloon.
5) The cladding balloon may burst, leading to dispersal of fuel in the coolant.
Summary of project
• A post-transient LOCA test rod has been examined with the GET instrument for the first time.
• Analysis methods for quantitative tomographic reconstruction of the fuel distribution has been developed. (Requires correction for strong and nonhomogeneous radiation attenuation in the object).
• The packing fraction has been evaluated in the balloon region of the test rod.
• Uncertainty estimation has been performed based on Monte Carlo calculations.
Analysis
• Quantitative reconstruction requires that the radiation attenuation in fuel and rig structures are accounted for.
• Rig structures as well as the fragmented and relocated fuel need to be localized. • UPPREC, a custom-made reconstruction
tool for tomography of nuclear fuel, has been used.
– Attenuating material locations are obtained from naïve reconstructions
– Rig structures identified using template matching in reconstructions of activation products
– Special module developed for disturbed fuel attenuation correction
Test rig structure localization
• Structures of the test rig were identified, using
– Activation products – Shadow effect
• Objects localized using template matching algorithm with UPPREC.
7 Reconstruction of 51Cr (320 keV )
Fuel reconstruction
• Preliminary (naïve) reconstruction is used to map the attenuation
distribution.
• The transmission through the fuel from
each pixel of the reconstruction to each detector position is calculated.
(~100 million lines of sight)
• After accounting for attenuation in fuel and structures, a final reconstruction is performed.
8
Left: the reconstructed emission intensity distribution of 662 keV gamma rays. Right: The calculated transmission of the 662 keV gamma radiation from each pixel to the detector (demonstrated at 0° rotation, i.e., detector is located to the right side). In both images, black corresponds to zero and white is set to the
Small collimator reconstructions. Several fragments and/or agglomerations of fragments fill the balloon. In the reference locations of the top of the rod, pellets appear solid and undamaged.
Some results of the fuel reconstruction
1-2 3-5 6-9 1 2 3 4 5 6 7 8 9
Packing fractions
The packing fraction has been estimated by combining
– The activity determined by tomography
– The cladding extension from photography
Packing fraction
Uncertainty evaluation
Random fuel distributions have been modeled using MCNP to estimate the accuracy and precision of the self-attenuation correction procedure.
RMS Error = 2.6 percent units (i.e. currently not the dominating uncertainty.)
Conclusions and outlook
• First quantitative GET reconstructions of relocated fuel in LOCA test rods have been performed.
• Attenuation in fuel and rig structures can be accounted for.
• Packing fraction measurements in the test-rod balloon have been
enabled.
• Measurements can be performed close to the reactor, with the fuel
distribution maintained as conserved as possible.
• Tomographic assessment is planned for the next LOCA test rod.
– We access the fuel next week!
Thank you for listening!
Comparison with naïve solution
Method RMS Error [percent units] Naïve solution 5.1 Proposed method 2.6 One extra iteration 1.9Collimators
Axial gamma scan (662 keV) with indications of the locations selected for tomographic inspection. The upper plot shows the locations inspected with the small
collimator and the lower plot shows the location inspected with large collimator.
Two collimators used 1x22 mm & 1x2 mm
Reconstructions using a large collimator
. Quantitative reconstructions using the large collimator showing the 662 keV emission intensity
1 2 3 4 5 6 7 1 2 3 4 5 6 7