Session: 11-01: Nuclear Power NDE

Paper Number: 139527

139527 - An Ultrasonic Approach to Identify In-Core Reactor Fuel for Safeguards 

Abstract: 

Operators directly sample fuel at their nuclear facilities on a routine basis, typically annually, for physical inventory verification. The reduced access to the cores in many small modular reactor designs could interfere with this practice. To overcome the issue, an ultrasonic system is proposed for continuous identification of nuclear fuel assemblies within an operating sealed or long-life core in a liquid-metal-cooled reactor. The process ultrasonically reads a series of notches of various depths, and subsequently translates them into fuel assembly identification numbers (IDs) using an encoding scheme. 
This work considers affixing brackets atop the fuel assemblies in an example lead-cooled reactor design that have up to 12 grooves machined into them, at up to 13 different depths. Each indentation is interrogated with its own ultrasonic sensor. This configuration is sufficient for more than 2000 distinct fuel assembly IDs. However, it can be adjusted as required by varying the number of grooves and depths. Three encoding scenarios are studied and evaluated with respect to the maximum number of ultrasonic transducers that can fail without compromising the fuel assembly identification.
The first of these assumes the fuel orientation is known via other means and the notches are dedicated solely to a Reed–Solomon encoding of the assembly ID. If five distinct indentation depths are utilised, this approach allows full information recovery even with failure of three out of 12 transducers. A model of the encoding is shown, in which a notched annulus, affixed to the top of the fuel bundle, is scanned in immersion through molten lead. The configuration is simulated ultrasonically in CIVA. The results indicate that a notch depth resolution of at least 1 mm is possible with an ultrasonic transducer of approximately 5 MHz centre frequency and 50% bandwidth. Additionally, the setup shows minimal amplitude variation in the echoes received from indentations with depths between 2 and 10 mm. 
In the second scenario some of the notches are used to orientate the assembly and the remainder of the notches are used for fuel ID. In practice, the orientation notches are required to be different depths from the assembly ID notches. In this scheme, there is an additional mode of failure in the loss of all the orientation notches and any of the encoding notches. The procedure now needs at least seven distinct depths to obtain full information recovery with three out of 12 transducer failures.
The final scenario utilises cyclic encoding, such that the fuel assembly ID can be determined regardless of which way it is loaded into the reactor. The procedure can be realised via a Reed-Solomon cyclic encoding and allows full recovery even with failure of eight out of 12 transducers. However, it requires sufficiently high ultrasonic resolution to distinguish 13 distinct notch depths. This is achievable with conventional ultrasonics, but requires a sufficiently deep annulus to support larger notch depth steps, which remain reliably separable over time. Alternatively, polynomial cyclic encoders can be optimised with respect to the number of depths. For 12 notches with five distinct depths, up to three transducer failures would not prohibit retrieving the full assembly ID. This solution matches the recoverability of the first scheme, without any prior information requirements. 
This work considered a liquid-metal-cooled reactor with a long-life core that could benefit from an ultrasonic system to identify its fuel on demand, for safeguards verification purposes. A range of encoding schemes were outlined, which convert the ultrasonic time-of-flight readings from notches etched on the fuel assemblies into identification numbers. Generally, as the number of distinct notch depths increased, the encodings became more resilient to random transducer failure. The study was undertaken solely in simulation. The long-term deployment of ultrasonic transducers in the high temperature and radiation environment of a reactor core has not yet been considered. Nevertheless, advancements in piezoelectric and magnetostrictive materials could allow the system to be brought to implementation testing.

Presenting Author: Rosen Rachev Canadian Nuclear Laboratories

Presenting Author Biography: Rosen Rachev is a R&D Scientist with over 7 years of experience in NDE for the petroleum and nuclear industries. His areas of expertise are in advanced ultrasonic array data processing, including ultrasonic signals modelling, component geometry reconstruction, phased array imaging and defect characterisation. He received a Master of Engineering degree in Mechanical Engineering (2016) and a Doctor of Engineering in Non-Destructive Evaluation (2021) from the University of Bristol.

Authors:

Bryan Van Der Ende Canadian Nuclear Laboratories
Mark Stringer Canadian Nuclear Laboratories
Mark Luloff Canadian Nuclear Laboratories
Rosen Rachev Canadian Nuclear Laboratories