Looking into plant matters

Manufacturers are benefiting from dramatic improvements in key performance indicators as the science of asset management undergoes a transformation in breadth and depth. Not only are our ideas changing from maintenance being about a delayed reaction to failure to instead a key instrument in the pursuit of business improvement, but it is also taking advantage of newer technologies to gather data from parts of the asset base that other systems can't reach. As Matthew Prior, business development director at systems integration firm Rex, which serves the oil and chemical industries, says: "We've moved from reactive maintenance to preventative maintenance, and the move is now towards predictive maintenance." However, while the theory is great, practice might be different. Durand Glass, for example, found its new maintenance management system a boon – but the more sophisticated benefits weren't so easily attainable. The firm employs 1,000 and operates round the clock making tableware: glasses, plates and mugs. Keeping machines running is vital, explains Fabien Klimza, project engineer and maintenance foreman: "You have to keep the furnace running – you only stop it once every seven years to rebuild it. So if a machine is down, the furnace keeps on making glass and you have to dump it in the basement. It does get recycled but there's a cost attached and you are losing precious production time." Maintenance at Durand used to be done using a combination of manual and proprietary systems. "Only administrative staff had access to the works orders – not the manufacturing or maintenance personnel – and the system wasn't integrated with either inventory or purchasing," says Klimza. "We could track the employee time, and hence the labour cost of a job, but were not able to request parts from the storeroom or purchase them from suppliers." Durand set three goals for a new system: to track works orders, whether for preventative or emergency maintenance; to track spare parts inventory; and to issue purchase requisitions with electronic approvals – workflow. "These three requirements are intertwined: if you want to make a preventative maintenance action you need to know what spares you have, and you may need to buy some so you need to make purchase requisitions. And you need to track both the spare parts and the labour needed to do the repair." After some research, Durand Glass decided on Maximo from MRO Software. "It did save administrative work, but beyond that one of the main benefits is to have a system that's accessible to everyone on the site," says Klimza. "The mechanics can directly enter or research any information they may need either on the stock level of any part, or if they had a breakdown they could put information in about what they did to repair the problem so the next engineer to work on the machine can see maintenance history." So far, so good. But what about the goals of analysing historical information to move towards better managed preventative maintenance, and then to predictive? "For that you need historic data," says Klimza. "You need a history of breakdown information entered into the system before you can see how often you get this type of breakdown and address your preventative maintenance frequencies." Give it time So the first problem is time – waiting for enough data to make the analysis worthwhile. And this presupposes that data is being collected, and that it is sufficiently comprehensive. "It does take some time and effort to put the data in, and it takes time for people to understand the value of putting in the information," observes Klimza. That's not surprising, when the real benefit could be some months or years down the line. Further, according to Klimza, "It also takes some effort to analyse the information in the system. It's not always easy. You create a lot of works orders and it's not easy to understand how to analyse them." Another company with unstoppable raw material is the Faccenda Group, one of the UK's leading suppliers of fresh chicken to supermarket chains. Privately-owned Faccenda operates from six sites in the UK, supplying large supermarket chains and restaurants with whole chickens, breasts and portions. Faccenda has complete control over the process, from the farms where the chickens are reared – and the associated feedmills – through to processing. Chickens keep on growing, and have to be processed before they get too big for the machines, so downtime can be seriously expensive. Previously, the company used a mix of manual and computer systems which, while they did the job, couldn't deliver the benefit of consolidating and analysing data across its sites: "We needed a system that would enable us to review breakdown reasons and introduce planned maintenance schedules to improve availability. When you're processing over 2m birds a week, downtime can be costly," says Phil Bevan, group engineering manager. We're running close to 24/7 and we have to keep the process going." Faccenda went for MVI Technology's plant maintenance and analysis system, and it's gone well. "We've now identified the main reasons for downtime," says Bevan. "We're doing more preventative maintenance – we've changed the shift patterns at some of our sites so that maintenance is scheduled when the machines are idle. Roll-out is now well under way at the second site." Key to Faccenda's decision to buy this system was user-friendliness, the point being that if people can't put the information in readily, it won't be there to be exploited. But the value of information also increases with reach – the number of people who have access to it. And traditionally, maintenance management systems have been big monoliths that belong to engineering. Today, however, few can afford for maintenance to be seen as something distinct from production. Not only does production clearly need to know what's happening in terms of the maintenance schedule, but there's another side, and it's to do with a gradual change in the way maintenance is being viewed. "Agility is a browser-based system where views can be put together in a distinct format for any target user population, for example for production," explains David Hipkin, managing director of systems developer SoftSols. Why? "Because production can easily and quickly make a valuable contribution to the engineering teams if they are logging data in a format that the engineering department can assimilate, and then take appropriate action on." Conventionally, machine condition data gets communicated electronically as a result of process automation. But, says Rex's Prior: "In any manufacturing facility only a certain amount of asset is automated." Even in the performance-conscious oil and gas, petrochem and chemical industries, where a day's lost production is measured in millions of dollars, there are still between 40 and 70% of assets that are not automated. "And the big area of opportunity that some of the oil majors are now talking about is operator-drive reliability." Meaning what? "It's the operator in the field who is looking at the asset on a daily basis, and is in the front line of asset integrity," continues Prior. "With technology like wireless environments supporting data collection and inspection routines using hand-held devices, you can combine that data with what's coming from the automation and control system. Everyone tends to think of asset management in terms of process automation systems from the likes of Emerson, Honeywell and so on, but one of the big areas of opportunity is the data that doesn't come out of those systems." If that's not enough, some manufacturers have woken up to the realisation that assets, and indeed daily expenditure, don't stop at production machinery. Some are getting even more benefit by collecting data they didn't traditionally collect, in areas like energy and building management – so expanding their definitions of assets, productivity, and thus also maintenance.