Shopfloor data is increasingly being recognised as key to everything from continuous improvement to business agility. Brian Tinham looks at how to get your plant assets talking’
Manufacturers that collect shopfloor data, and integrate it with their management systems and processes, are finding their new-found visibility invaluable – sharpening their focus on what can otherwise be the black hole of production. It’s helping them to resolve production problems more quickly, be more responsive to customer requirements, provide accurate delivery dates, improve maintenance operations, transform overall equipment effectiveness (DEE), even get better job costing. The list goes on.
And there’s more than just anecdotal evidence for that assertion. Earlier this year, a survey by the Manufacturing Execution Systems Association demonstrated conclusively that manufacturers sharing KPls between operations and finance show clear advantages over those that don’t. Its research, (sponsored by GE Fanuc, IBM, Rockwell and Siemens, among others) revealed that winning manufacturers use a ‘metrics framework’ that links production to the business, using shopfloor automated data collection.
But it also found that only a fraction of manufacturers have these factory connections. Indeed, just 3% of respondents reported ‘very effective’ links between operations’ KPIs and business metrics. Which is worrying: “What we’re talking about is manufacturers’ survival,” insists Julie Fraser, principal of analyst Industry Directions, which undertook the study. “If operations and finance aren’t on the same page, you have a company at cross-purposes. Most manufacturers can’t afford to be in that position today.”
Which raises the question, if it’s that important, why aren’t more of us doing it? Well, it seems there are various reasons, but three are most commonly cited. First, plant assets aren’t connected consistently to any one system, so there are technical problems with protocols, languages and even available communication ports. Second, it requires analysis, investment and dialogue that crosses departmental and budget-holder boundaries. And third, we’ve invested in business/finance-orientated systems (such as ERP) and expect them to deliver on this promise. Perhaps we should add a fourth: many operations managers are aware of their lack of detailed factory insight, but they don’t know whereto start solving the problem – and democracy by decibels doesn’t help.
Some of this is fact, some just perception and some downright fallacy, but we just have to get over it. So let’s set the record straight on the IT side. The first point to make is that connectivity really isn’t the big deal some might have you believe. Protocol converters and connectors can handle pretty much anything, old or new, while developments with fieldbus and the ‘higher’ digital plant monitoring and control networks are making integration easier and cheaper, both to install and maintain. What’s more, wireless options are growing almost by the day.
Second, shopfloor data capture (SFDC) and barcode swipe terminals, that not only interface directly with machines, but work centres, whole lines and (most important) people, have also been around for years. They may not be the solution itself, but they’re a key step to getting reliable, consistent and accurate factory feedback without paper and without adding waste – and potentially also providing operator direction and help.
And third, there are now useful manufacturing execution systems (MESs) and production management tools and systems out there. Admittedly, ‘MES’ remains an over¬used and under-defined term, but nevertheless, the newer wave of largely configurable, modular packages – some very industry-specific – do seem to be getting to the heart of factory communication and automation issues.
So yes, there are different ways of going about this. And, yes, there are the political hurdles, as well as the ERP scepticism to overcome. But there are now systems, some neither expensive nor long-winded to implement, that are transforming operations and business alike, from the bottom-up.
Let’s turn to data collection terminals and barcoding, taking the experience of Durham ¬based Dyer Engineering. The firm makes low volumes of high quality components for the offshore industry, medical sector, defence companies and engine manufacturers, and everything is make-to-order. Traceability is important to its customers, and not just in the defence sector – so that was one driver for its SFDC project. But managing director Paul Dyer (above) also recognised that being able to track material and components through production in near real time was key to understanding issues and making improvements around, for example, bottlenecks. As the firm grew, all that became more important. Says operations manager Margaret Grievson: “We’ve been in business for 30 years and for much of that time we were able to see every component on the floor. But as the business grew we couldn’t get the same view. There are so many parts – about 40,000 a month – that we lost vision of where they were, and couldn’t maintain traceability.”
The company implemented barcode and swipe card readers throughout the shopfloor and logistics operation as part of its Exel ERP implementation in 2004, and in the last couple of years the organisation has been transformed. Now, each works order has barcode for every step and that’s read at the point of process. Each operator and every workstation, indeed every plant unit, also has a barcode, and that too is logged, providing comprehensive insight into everything from start to finish. And the same applies to goods-in, purchasing, despatch …
“We now have, in effect, a live balance sheet and a live profit and loss account,” says Dyer. “The detail across the whole company – ¬all activities, all individuals and all aspects – is so accurate, so real-time, that we can see how the entire business is functioning at any time.” He says it’s difficult to quantify the benefits, but suggests that the firm could not have expanded without live data. “What it gives us is vision – real-time vision, where we are updated potentially every second, instead of waiting for historical data. This way, we can see blockages holding up a process and take a decision on investing to get it back on track. We’re able to see where a job is and whether it’s taking the right amount of time and feed that information into the balance sheet.”
And since it’s also linked into the time and attendance system, it also links productivity to people’s pockets. “The emphasis is on rewarding our workers for their hard work,” says Dyer. “If they exceed targets, they will be rewarded. If someone is struggling, we’ll have the means to know why, and use that to help the person concerned to improve.”
Getting there wasn’t easy though. Dyer indicates that it’s only in the last couple of years that everything has come together. “The most difficuLt part was instilling the discipline into the workforce to swipe on and off operations,” he says. “We had silly scenarios where some of them thought that if they didn’t swipe on and off they wouldn’t be noticed – but they soon found out when it impacted their remuneration!”.
An important point here is that Dyer Engineering opted to capture data from all of its production and beyond – and, for many, that’s probably overkilL. But the issue then is what to capture, integrate and/or display, and what to omit? And there’s the potential for problems. Mark Sutcliffe of CDC Software (formerly MVI Technologies) says: “The danger is someone senior starts with the ‘business
pai n’ and then gets frustrated because they can’t find the causes. So they think ‘I’ll gather data’, and that’s when MES, OEE and dashboards appear and they’re drawn into a project.”
Which is one way of saying you don’t know what you don’t know, so even the best of structured approaches can fall prey to popular initiatives that might or might not be the right answer. Best advice is to go for the bottom-up, holistic approach recommended by those who take their MES very seriously¬ the likes of Siemens, Rockwell, ABB, Mestec (which has been working closely with Oracle) and again CDC Software.
Siemens’ Richard Porter puts it thus: “I would Look at the process overall- talk to the production manager, the shift managers and operators, and try to understand where their problems are. You might find asset utilisation is poor but no one is sure why, so you will end up putting in an DEE solution with a historian and a plant performance analyser. Then you’ll get useful context, such as which product, in batch ID, which shift, which production line and so on.”
Sutcliffe agrees, but makes the point that shopfloor data can also be the key to all sorts of other business information – such as finding out the hidden profit or loss by SKU or product mix, or capacity opportunities, training needs, supplier issues, plant investment requirements and even issues between maintenance and operations. He suggests thinking about the factory process stages, insight into which would make a significant difference to all that – and then the equipment, operations and people that map to them.
“OEE might turn out to be as useful as a chocolate teapot, if what they really need is something that helps them understand the reasons for stoppages, the context, the differences between products and shifts. Something that collects relevant plant data, allows operators to indicate reasons for downtime and provides the analysis in real time, solves plant problems very quickly. It also engages people, gets them competitive and makes continuous improvement a natural thing they’re motivated to do.
“It’s what Toyota discovered a long time ago. Continuous improvement isn’t just about kanbans; it’s not just emulating processes; it’s also people – and Toyota has 10,000 people who think about doing their jobs better every day. What a good MES does is institutionalise good practice, let people see themselves with measures they can understand and make that information visible.”
Greencore, one of Europe’s largest food businesses, with manufacturing plants across the UK and Europe, swears by CDC Factory. The company wanted to integrate real-time production operations with quality assurance processes and improvement campaigns, but have it driven by the factory floor. Says Greencore’s operations director David Gallagher: “We wanted to try and reverse the usual relationship, so that pressure to make changes is being driven upwards throughout the organization. We had a very strong opinion that it had to have a holistic approach – it had to connect the shopfloor to what the management system was doing.”
Gallagher chose Greencore’s cake and desserts-making facility to pilot the approach. The site already had access to long-range operational performance data, but as he puts it: “What we lacked was information for people on the shopfloor that can affect the next five minutes.”
Go-live happened after just eight weeks, and the results speak for themselves, in terms of time and money_ In one area, changeover costs and times were reduced dramatically. Not only did the production and scheduling teams quantify the benefits of scheduling similarly-coloured SKUs for production in the same period, but they identified low-cost opportunities to save wash-down time by modifying conveyors for self-cleaning. Elsewhere, one ofthe zones in the dessert manufacturing plant was found to be suffering recurring downtime. CDC Factory established the primary cause as ‘waiting for mix’. Installing an additional mixing vessel with a cross-feed pipe to the depositor meant the mixture could be prepared continuously and fed in to the production line without downtime.
“We were very clear with the workforce that the intention was to allow them to come up with new ways of working,” says Gallagher. “We wanted them to spend less time just coping and more time improving. They have embraced the empowerment that CDC Factory gives them … Previously I would be a prime mover – now it’s much more ‘Tell me what the options are, yes I think you’ve chosen the right one, go ahead and do it’.
“It’s the people who work with a system that make things happen.”.