Keeping a close eye on nuclear waste11 November 2021
A new inspection system should increase the speed and effectiveness of visual inspection of waste at Sellafield
The UK’s National Physical Laboratory (NPL) is the country’s national metrology resource. It is a public corporation owned by the Department of Business, Energy and Industrial Strategy (BEIS), and it develops and maintains the national primary measurement standards, provides a national measurement infrastructure and delivers the UK’s ‘Measurement Strategy’ on behalf of BEIS. It aims to use science and engineering to provide the measurement capability the country needs.
Dr Nick McCormick, NPL’s principal research scientist, says, “At NPL we are interested in developing techniques to make measurements less subjective and to minimise human variability. The automation ensures the inspector can concentrate on areas of potential concern and use their skills to efficiently make an accurate assessment of conditions.”
This year NPL has deployed a ‘High Accuracy Inspection System’ (HAIS) to Sellafield Ltd to carry out regular inspections of nuclear waste stores.
The HAIS was conceived by NPL, and it uses digital image correlation (DIC) technology to analyse image sets and quantify changes over time. This imaging technique has previously been used in sectors including rail, aerospace and oil & gas, but NPL says it is particularly well suited to monitor the integrity and conditions of different materials and using it to examine evolution of nuclear materials allows extremely small changes to be detected more rapidly than using traditional methods.
NPL explains that DIC examines two images — before and after — at the pixel level by creating small subset images. These images are compared for cumulative differences after applying x/y displacement, strain, shear and exposure differences. Using optimised techniques, the best fit for these parameters are found for each small subset of the image and hence a fulfilled measurement of displacement, strain shear and correlation coefficient are generated. NPL adds that one of the challenges of DIC measurement is controlling the position of the camera to reduce parallax effects and to provide sufficiently reproducible position and orientation of the camera to allow successful DIC for interrupted deployment.
DIC has been used since the mid 1980s, but it has only been within the last 15 years that cheap high-resolution imaging and appropriate computing resources have been available to make measurements within a sensible time. DIC is not widely used outside of the laboratory and it has been work at NPL that has identified the uses for this technology in diverse civil engineering structures like tunnels and nuclear stores.
The HAIS system is a precision robotic platform that allows high precision repeatable image capture with high quality images, so it can be used to inspect nuclear waste products in storage.
Addressing the problem
For many years, encapsulated waste products have been stored at the Sellafield Ltd site and over the next decade retrievals from legacy ponds and silos will generate more. As Sellafield Ltd evolves into a waste management and remediation site the range of waste products is also expected to increase.
It is vital, for safe storage, that a programme of inspection of the waste stores and their key properties is undertaken. However, with many thousands of packages located in engineered stores, it is time-consuming to extract and inspect them all. Therefore, in-situ measurements and deployment techniques are required to demonstrate control, allow appropriate mitigating action when necessary, and reduce dose to workers who currently provide ad-hoc measurement capability.
HAIS can carry out regular inspections up to 16m deep into low-level waste stores. HAIS deploys a camera vertically into an inspection port and takes a series of images inside the waste storage at pre-determined points.
The HAIS system used at Sellafield has been designed for this particular environment, where WIFI is not possible and hence it is cable operated, using digital transmission between components. It provides its own illumination as the environment is unlit and has to be able to traverse up to 16 metres down into an environment whilst capturing high quality images.
During the inspection the HAIS uses DIC to analyse previous image sets and quantifies changes over time, including corrosion, movement, vibration and dirt or water ingress.
The changes are anything that changes the appearance of an image. That can be displacement or surface deformation, but also additions and subtractions that occur between images. This is what allows things like corrosion, cracking and loss of components to also be measured.
Areas of concern can then be highlighted and monitored, or a closer inspection carried out since the position is accurately recorded.
Human in the loop
The human operator is an important part of the HAIS system. The system has been developed to identify areas of significant change. This augments an inspection by highlighting only those areas of significant change and — most importantly — not showing areas of little or no change to an operator.
NPL says this means that the system does what a computer does best: identifying change. It then lets the human operator make a decision as to the nature and significance of the change. This allows a seamless connection with current safety cases for inspection and still requires the ability of a skilled inspector to sign-off the case, but without the boredom of surveying areas of no change.
It could be possible for alternative techniques using machine learning to be used in place of the human operator, but in practice they require too much data on too many different types of defects to allow reliable detection and confidence in the black-box decision making.
There are many benefits of deploying HAIS to monitor and inspect waste stores, including gaining a greater understanding of the evolution of the nuclear materials stored and how it affects long-term safe storage. For example, measuring the properties of materials in-situ allows for a greater understanding of how much heat is being generated by the material and what the storage system needs to tolerate it. These measurements will ultimately underpin the storage strategy.
For existing stores, HAIS is able to obtain a better understanding of how store environments can change under a range of scenarios for different waste packages. Having accurate in-situ measurements increases the predictability of the environment and therefore helps to infer waste behaviour. The system also provides the tools for monitoring the environment to provide an early indication of deviation from the predicted environment, including temperature, humidity and chloride. Understanding the long-term optimum storage conditions for a range of products enables the system to provide information that could be vital in influencing new store design and improve existing store conditions.
Using the wider history
The history and experience of using DIC in other civil engineering structures helped guide NPL in developing and testing the HAIS system. This included road and rail structures, oil and gas plant but some measurements were much more diverse — NPL says it has even used DIC techniques to measure setting chocolate for factory production lines.
Clearly there are also different options for deploying HAIS, for example continuous monitoring or regular inspection, depending on the nature of the objects being monitored.
An alternative, fixed cameras that can make measurements frequently, has been demonstrated by NPL in regard to CCTV used for monitoring assets belonging to the UK’s Environment Agency. NPL is also developing DIC applications that capture data on railway tunnel walls whilst moving along the track. Its most recent development is using a camera within a mobile phone for repeatable measurements, this could be used for monitoring widely distributed assets like bulkheads in ships.
NPL says both types of strategy are possible but generally, for lots of repeated objects that need to be monitored intermittently as at Sellafield, a single device that can be moved around is preferred.
What are the benefits of the HAIS system? Using DIC is expected to increase efficiency across the Sellafield Ltd site by enabling very small changes to be detected and far sooner than using traditional manual inspection techniques. The inspection and measurement techniques, enabled through HAIS, uses in-situ automated technologies in dark store environments where traditional communication channels or power sources are absent and allows the detection of signs of unexpected degradation.
In addition, NPL points to the objective nature of the measurement, and the ability to remotely record data, which can then be archived. When the data has been analysed, different inspectors can contribute improving the quality of the measurement and removing subjective and systematic biases. The system also more effectively uses the time of a trained inspector and reduces fatigue.
Dr Robert Bernard, Senior Technology Manager, Sellafield Ltd, said: “Sellafield Ltd are happy to be working on introducing new technology to support our mission of reducing the risk of storage of the special nuclear material we’re entrusted with by the nation. Having scientific organisations such as NPL partnering in our work ensures we’re at the cutting edge of deploying practical technology to maintain safe storage.”
Keep up to date with technical features
This article was first published in the October issue of Nuclear Engineering International magazine. Sign up here to receive a sample copy of our latest digital issue.