Modelling to save Sellafield!

At the February Sci-Bar meeting, Alex Brown, a chemist from the National Nuclear Laboratory in Warrington delivered an enthusiastic presentation on in silico approaches [the current description of computer based data analysis] to managing the challenges of cleaning up Britain's first nuclear plant at Sellafield. Using a set of images and a time-lapse video, Alex took us through the acronym-filled world of the former Windscale site, which grew out of the earlier Royal Ordnace works, to take advantage of the early stage developments that underpinned the nuclear weapons industry. The various components of the plant that have been built since the World War II, comprise reactor towers, fuel waste ponds and a range of waste storage areas. These are all there to deal with the the original purpose of power generation via the Magnox system, and the safe storage and disposal of on site waste. As Alex reminded us, Sellafield also handles waste brought in by rail from other plants around the country. It is likely that complete decommissioning of Sellafield will take up to 100 years from today. The inevitable "fall out" from an industry that incorporated a large number of empirical process decisions, as this embryonic industry developed, is a mixture of decaying radioactive waste in formulations that are not entirely understood. As with all complex problems in Science and Engineering, the application of computational modelling is being increasingly deployed in order to manage the clean up processes, with a view to developing the safest, most cost efficient and quickest solution to restoring Sellafield to its former rural beauty! Such problems are of course very familiar to the citizens of Widnes and Runcorn, who live alongside similar consequences of the early Chemical Industry in these towns! The ever present dichotomy of job creation and technical advance versus health and environmental risk!

Alex used a specific example of the pH related problems encountered in the management of settled sludge in a particular storage "tank". Acidification caused by aeration of the sludge tank, had an important influence of the solubility of Brucite, one of the Magnesium containing components of nuclear waste residues from Magnox reactors. Alex took us through an explanation of the use of a standard equilibrium equation for the solubilisation of Brucite at a range of pHs, starting from above 11, down to just above 10. By calculating the predicted solubility of Brucite in solutions of different pHs, the aim was to predict the Brucite concentration over a period in the on-site container, using a set of known temperatures and pressures. However, the modelling team had some concerns over elements of uncertainty surrounding the tank dimensions and the mixing parameters. Unfortunately, the model showed considerable deviation from the observed Brucite behaviour. 

By adding magnetite into a set of simulation experiments, a closer fit to the observed solubility characteristics in the tank was obtained. As the pH drops from its starting alkaline value as dissolved, atmospheric CO2 acidifies the liquor in the sludge tank, there is a concomitant impact on the equilibrium position of the liquid versus solid phase of Brucite in the waste mixture. It seems to me that since this is an accepted outcome of the plant configuration (ie aeration cannot be conveniently contained), the model requires incorporation of the multiple pH dependent events that will collectively determine the distribution of ionised Brucite in its dissolved and precipitated phases in the mixture. It was good to hear that Alex and his colleagues are now developing more sophisticated models in order to address such issues.

Alex, was clearly a passionate advocate of modelling approaches to helping manage such processes,  and he provided further examples of empirical curve fitting procedures in order to deal with client generated data (which raised an eyebrow in the audience!). Irrespective of the physical basis of such curve fitting equations, providing they provide the client with a robust means of predicting behaviour, they are valuable. Industrial protocols can be developed from such modelling to manage processes such as nuclear waste removal and, in more progressive examples, to help design new generation nuclear reactor plants. Alex's colleagues were in negotiations with Chinese organisations fro this very reason.

As is the norm at the Widnes SciBar, the audience contained a sprinkling of seasoned veterans of the nuclear industry, who shared some of their fundamental knowledge and experience, which added colour to the talk. Alex also fielded several questions about the value of modelling, the time and computing power needed, and Bob invited Alex to comment on the future challenges and opportunities for Nuclear Power. There were also some detailed questions about gaseous waste issues, the replacement of Magnesium casing with Zirconium and several more general questions about site access and visitors. Bob, it might be an opportunity for a SciBar coach trip! In the absence of an immediate trip, "Britain’s Nuclear Secrets – Inside Sellafield" with Jim Al-Khalili which was on the BBC last year., isn't currently available via the BBC, but it is here:- www.dailymotion.com/video/x3orm62

Thanks again to Alex and all those who attended: another full house at the Hillcrest Hotel. The next talk will be announced by Bob, and looks at the future of technology through the eyes of Star Trek's creator, Gene Rodenberry! And I leave below a link for those wishing to swell the ranks of the British Science Association. Personally, why wouldn't you!

The British Science Association can be accessed here. 

Dave Hornby and Bob Roach