At some stage in the decommissioning of nuclear installations (e.g., reactors and gloveboxes), it is inevitable that large metal structures (e.g., reactor vessels, gloveboxes, structural components) need to be cut into smaller pieces to be packed in containers, for temporary storage or permanent disposal or simply for transportation. Both cutting and packing cost money. Usually, efficiency or cost-saving in one is achieved at the expense of the other. So a trade-off is always required. For example, a heavily radiation-protected human operator goes into a hot cell to take down an installation. Because of radiation exposure limit, there is usually not much time for deliberation.
The decision is by and large made there and then (i.e., at the discretion of the human operator who is at the scene), unless (on a rare occasion that) a prior knowledge about the hot cell happens to be accurate enough that planning/schedule done beforehand can actually be relied on. Once a cut is made, there is no going back, i.e., little or no margin for correction. There is an on-going project, supported by NDA/Innovate UK, to demonstrate the feasibility of using robotics to carry out the structure scanning, radiation mapping, dismantling and packaging tasks. The dismantling and packing is guided by a cutting/packing simulation model developed at Leeds.
The use of robotics can in principle remove the constraint relating to radiation exposure time (and hence allowing for longer operating time). While the cutting/packing software model at the moment allows a human operator to perform cutting/packing trials entirely on a computer, there is not much machine intelligence built in to search for ‘optimal’ schedule automatically. Brute force trial of every possibility is a simple idea, but even a small packing, the number of trials would quickly spiral out of control. This project aims to work out and introduce an Artificial Intelligence approach to the operation, whereby the machines can on their own decide which part of the structure to scan/rescan, where to cut and how to pack, and in what sequence, etc, in order to ensure the total cost is minimised.
Short of brute force trials, there is an intelligent and scientific way to optimise, under various constraints, the cutting/packing process for overall cost saving, assuming accurate structural model and radiation mapping are available (from in-situ, real-time, robotically operated scanning).
Aim: To find, implement and verify AI solutions to optimisation of dismantling/packaging process in nuclear decommissioning.
• Search and review relevant optimization routes. • Develop them such that machine autonomous operation and decision-making is possible. • Demonstrate feasibility through simulations. • If possible, implement and demonstrate at lab scale using real robotics.
Methodology and Approach
In shipping, methods to optimise stacking of freight containers do exist. The challenges for nuclear decommissioning are that (1) the objects are rarely of a regular shape and/or of a uniform size, and (2) the objects are created as a result of cutting (i.e., the feed is not pre-fixed).