Robots get smarter to help with decommissioning

A laser-cutting robotic snake has been deployed on the highly-radioactive core of the redundant Dragon reactor at Winfrith in the UK. Meanwhile, machine learning algorithms to increase capabilities in several crucial areas of nuclear robotics - including waste handling, cell decommissioning and site monitoring with mobile robots - are to be created by a team of computer scientists from the University of Lincoln.

A laser-cutting robotic snake has been deployed on the highly-radioactive core of the redundant Dragon reactor at Winfrith in the UK. Meanwhile, machine learning algorithms to increase capabilities in several crucial areas of nuclear robotics - including waste handling, cell decommissioning and site monitoring with mobile robots - are to be created by a team of computer scientists from the University of Lincoln.

The LaserSnake in action at Dragon (Image: NDA)

Magnox, which manages the Winfrith site in Dorset for the Nuclear Decommissioning Authority (NDA), called in contractors OC Robotics to help remove a 40cm-diameter vessel, known as the purge gas pre-cooler (PGPC), attached to the core of the Dragon reactor. One end of the vessel was attached to the core in the high-radiation area behind the reactor's concrete shielding and several steel plates, while the other end extended outside the shielding.

The LaserSnake long-reach snake-arm robot with laser cutting optics was developed by OC Robotics and TWI with funding from the NDA. It has previously been deployed at the Sellafield site in Cumbria.

The thick pipework, complex layout of the PGPC and limited access meant it was necessary to prepare two mock-ups to allow comprehensive testing and trial runs to take place. The LaserSnake was inserted through a narrow hole in the 3-meter-thick concrete shielding around Dragon's core and the actual operation to cut away the PGPC took less than three hours.

"The difficult environment of the external core of the Dragon reactor was an ideal challenge to show the full capabilities of laser-cutting technology and snake-arm robots," said OC Robotics' Adam Mallion. "Cutting something as thick as the 400mm PGPC with its complex internal geometry had never been attempted before."

Andy Philps, Magnox senior project manager, said: "We believe this is the first time that laser-cutting technology has been deployed directly on the core of a nuclear reactor. The ability of the LaserSnake to carry out 'keyhole surgery' on the reactor core meant that the work could be carried out using existing protective shielding."

Melanie Brownridge, head of technology at the NDA, commented: "This is an excellent example of how early NDA R&D funding support enabled the technology to grow from an exploration of whether laser-cutting could actually be adapted for nuclear into a system that, with further funding and collaborative working, is now mature and being successfully deployed on a number of our sites."

Winfrith was home to nine unique reactors and associated facilities before decommissioning work began in the 1990s. Today, only two - the prototype steam generating heavy water reactor and the Dragon prototype high-temperature gas-cooled reactor - remain. All that now remains of the Dragon plant, which operated between 1964 and 1975, is the reactor core contained in a pressure vessel surrounded by the concrete shielding, seven steel containment plates and an outer containment building. The reactor core is scheduled to be removed by 2021 and the remaining facility demolished by 2022. Large areas of the Winfrith site have already been delicensed and freed for future development.

Making nuclear robots smarter

Researchers at the University of Lincoln are developing artificial intelligence (AI) systems to enable self-learning robots to be deployed in hazardous nuclear sites.

Machine learning is an application of AI which enables systems to collect date and use it to form automated decision-making and make improvements based on experience without being explicitly programmed. The team at Lincoln will create algorithms for vision-guided robot grasping, manipulation and cutting, mobile robot navigation, and outdoor mapping and navigation. The aim is to build systems which can use machine learning to adapt to the unique conditions of nuclear sites, including locations contaminated by radiation.

A dedicated bimanual robot arm which will be mounted on a mobile platform is being developed. It will be operated using shared autonomy – where the machine is able to operate autonomously while still having humans as key decision makers – or via remote control. The team will also investigate the potential of augmented reality in the field of nuclear robotics.

The project has secured GBP1.1 million (USD1.5 million) in grant funding from the UK's Engineering and Physical Sciences Research Council. It is part of the National Centre for Nuclear Robotics (NCNR), which aims to develop advanced robotics and AI technologies for nuclear industry applications. The University of Lincoln was one of eight universities that founded the NCNR.

Project leader Gerhard Neumann said, "Clean-up and decommissioning of nuclear waste is one of the biggest challenges for our generation and the next, and the predicted costs are enormous: up to GBP200 billion over the next 100 years." He added that recent disaster situations, such as Fukushima, had shown the crucial importance of robotics technology for monitoring and intervention, which has been lacking up till now, "making our work even more vital".

Researched and written
by World Nuclear News