Finding what is left

Morson has been carrying out asset condition survey work for five years at Sellafield legacy ponds and silos. Having painted a picture of equipment on these shut down, but not-yet-decommissioned fuel reprocessing sites, site manager Sellafield Ltd can now understand better which assets are currently most at risk of failure.

“There is a radioactive inventory in those buildings, and they are in storage until the retrieval phase takes over and they are decommissioned. They are now in safe storage. Decommissioning of the first plant is still probably 20 years away. So those plants need continuous process function until they are decommissioned. You can’t walk away from it and expect it to take care of itself,” Morson project manager David Garratt tells NEI.

The problem is, Sellafield Ltd’s current maintenance programme is limited to assets critical to the safety case. “What we are doing is extending the level of maintenance that they carry out on safety-related work to other equipment that does not fall under its remit,” Garratt says.

“For example, if there is a water system, Sellafield will do regular maintenance of pumps and valves. The vast majority of the system, which is pipes, pipe supports and insulation, is unlikely to be looked at regularly. What we propose is, if they don’t look at it regularly, how can they ensure that the system will continue to function? If parts are corroding, how will they know that is happening? That is what this is about.”

The Morson team, a core of six engineers and surveyors, worked with site staff to understand the issues involved with assets in the buildings. The success of that maintenance process depended on gaining the trust and buy-in of the experts, according to Garratt: “They have the most insight into the way the plants operate, and their problems, and their maintenance, that senior managers would not have.”

“They tell us the condition of the plants, whether they have been maintained, what would happen if they were to fail, could failures create secondary problems, could plant failures potentially kill someone, are they near a watercourse; do plant staff understand where the plants are, where the assets are and what are the consequences of failure. The staff are best placed to answer questions.”

With this information, Morson whittled down a list of key maintenance criteria. This list now includes asset condition (whether they have been maintained); importance to the business case; future business needs; actual life vs design life; stakeholder interests (such as whether an asset cost might run over a budget); environmental consequences; risk to people; radiation safety. Each asset is broken down into questions, the answers to which are run through a computer to develop a risk-based score. Depending on the risk, the asset may be inspected every three or six months, or at a frequency ranging from yearly to once every five years.

By repeating quantitative observations over years, the Morson team expects to be able to understand trends in the assets’ decline to be able to predict residual life and plan maintenance in a proactive way.

Garratt explains how this might work in practice on a water line:

“Based on a design background, we can see where there is most likely to be wear on that component, and where it is most likely that water might start to penetrate the insulation and start to break it down, and we will target those areas. We will go and look at those areas every time and ultrasound the wall thickness; where it might have been 5mm originally, after five years it is down to 3mm, and we can see the rate of wear. We can build into documentation trigger levels so that before it reaches 2mm, it will need to be replaced within two years, and raise a project to replace it. Then there could be a second trigger if the measured thickness were 1.5mm that we need to do something within 12 months because the rate of decline is occurring faster than expected.”

Garratt admits that reading a trend out of a few data points might not be straightforward. “We have suggested an intelligent customer resource for the interpretation of data. That is a key part of this. Carrying out inspections relies on the interpretation of the data; otherwise they are worthless.”

Although the maintenance process was developed at Sellafield for Sellafield, it could be equally carried out across operating nuclear power stations, or even other industrial sites. “Most people think that nuclear power stations have a lot of exotic equipment,” Garratt says. “But in reality, most equipment for nuclear sites is the same as that at oil refineries and chemical plants elsewhere. There are exotic pieces of equipment, but the vast majority of equipment are pipe systems, ventilation systems, lifts, pumps and cranes.”