Automated sizing and classification of defects in CANDU pressure tubes

Abstract

Pressure tubes within CANDU reactors are subject to frequent ultrasonic non-destructive examination to identify and characterize any defects that pose a risk of initiating Delayed Hydride Cracking (DHC), a well-known problem that occurs in zirconium components that are subject to high mechanical and thermal stresses. The analysis of the ultrasonic data gathered from the pressure tubes is often a long and repetitive process on the critical path to the restart of the reactor. Motivations for developing an automated system include saving time on the critical path, minimising human subjectivity from the process and increasing repeatability of measurements. An automated system providing decision support to analysts also reduces the risk of operator fatigue by minimising time spent processing routine defects. This paper describes a novel system for the automated analysis of CANDU pressure tube inspection data. The system automates the entire initial examination process including the complex decision-making procedures by implementing an expert system with an associated rule-base. Results from this system are detailed, illustrating the location and characterisation of defects and key features within the pressure tube with a high degree of accuracy while increasing repeatability through the removal of subjectivity in measurements.

Evaluating Diffusion Models for the Automation of Ultrasonic Nondestructive Evaluation Data Analysis

“…Automated Data Analysis of Pressure Tubes, an end-to-end expert system that provides decision support and explainability for automated flaw characterization in UT NDE data taken from PT, is proposed in [13][14][15]. The system is based on deterministic rules and explainability is generated using a tree-based system developed with input from practitioners.…”

“…All referenced work shares a dependency on labelled data and out of distribution generalization. In the case of [13][14][15] labels and analysis are required in order to generate the set of rules on which the system operates. In the case of [17], labeled data are required to train the neural networks used to identify flaws.…”

“…The robustness of any system relying on labelled data depends on the amount of labeled data provided; more is considered better. The use of large datasets supports generalization in the type of deep learning systems trained in [17] and will allow rules-based systems, as in [13][14][15], to account for a wider variety of field conditions. The difficulty in terms of real-world applications of sufficient complexity is that field conditions may vary widely and in an unanticipated fashion, so that a dataset of any size cannot be guaranteed to provide the information required for sufficient generalization.…”

3-D visualization of AGR fuel channel bricks using Structure-from-Motion

“…120 the nuclear industry for a variety of tasks such as defect detection [1,7,8] and repair [9,10], accident response [11] and reactor decommissioning [12] across many reactor designs such as the AGR, the Canada Deuterium Uranium (CANDU) 125 reactor, Boiling Water Reactor (BWR) and the Pressurized Water Reactor (PWR). With each reactor having unique architectural characteristics, application-specific equipment is often developed to visually inspect a variation of life-limited compo-130 nents pertinent to each aforementioned reactor design.…”

“…A variation of RVI methodologies are deployed such as Ultrasonic tools to investigate weld joints [13] in the Reactor Pressure Vessel (RPV) and de-135 fects in CANDU pressure tubes [7], bespoke underwater robotic camera systems to investigate the PWR RPV [10] and lower core plates [9], pipe inspection robots for feeder pipes in the Pressurized Heavy Water Reactor (PHWR) and tethered inspection tools used within the AGR fuel channels and control rods [2,3,1]. The majority of visualisation techniques however are limited to using signal and 2-D image processing to evaluate components within the respective reactor designs, with 3-D visualization techniques only being recently deployed to evaluate Fukushima [11,12] to assess reactor station damage.…”