MES: a panacea for the plant pain?

What, within the 'four walls of manufacturing' are your real points of pain?" asks Matt Holland, product manager for Siemens' Simatic IT. Is it product quality, safety, regulatory compliance, OEE (overall equipment effectiveness), scheduling, maintenance, or disparate procedures across global facilities? All those are common issues that MES (manufacturing execution systems) can help alleviate. What we need to do is understand first that it can be done, then second put aside the IT, and third, look at the business of our manufacturing to prioritise the problems. That's what winning companies are doing today, says Holland. "I see [MES adoption] being much more operationally-driven now, with more focus on projects tangibly related to benefits," he says. Holland reckons it's all happening as companies are forced to change their technology culture. He sees it as being squeezed – both from the bottom up with "a more simplified architecture landscape on the control side", and from the top down, with ERP increasingly addressing plant and factory issues. "IT departments are having to give up some ownership," he says, "in the realisation that IT is moving down onto the shop floor, driven by manufacturing's needs." The bottom line is that the classic brick wall between business/finance IT and that in operations and engineering is being torn down thanks to open technology and "Microsoft coming to the shop floor" as he puts it. Why? Because companies are seeing the benefits of what amounts to a single underlying infrastructure in terms of gaining functionality and agility with low cost of implementation and support. That said, manufacturers need a vision. Dominic Malloy, global manufacturing solutions director at Rockwell Automation: "There must be a clearly defined three-to-five year business strategy. Identify the major obstacles to this strategy and the KPIs [key performance indicators] you need to check whether you are hitting the goals set." As John Bailey, managing director at Pantek, says, too few companies have gone down to this level of definition. They believe they want MES, but with something as broad scoped as this, they need to be clear about possibilities, priorities and deliverables. There are so many aspects to MES. "Most companies realise they need the MES 'glue' in the middle," believes Stuart Richards, consultancy manager at Pantek – the point here being to integrate business and plant data and get the benefits of that. MES then means you know what you're doing – meaning that real-time, hourly or daily shop floor information, for example, enables fact-based decisions at both levels. MES can then be about moving to managing exceptions, such as with material non-availability, machine downtime, rework or product recalls. As Malloy says: "It's the ability to react quickly." If you know how long it takes to make a product, the finite schedule and productive capacity of plant, you can deal with suppliers' and customers' requirements on the one hand, and your own issues on the other. A real-time scheduler tightly linked to the production floor, for example, tells you what you can make based on units actually available, down time, current run rates and so on. Then you can meet a key customer's rush order, gauging your ability to reschedule your manufacturing, and give firm commitment. "It works at both ends – you can be more responsive to the pull from the marketplace and more predictive to your suppliers," observes Malloy. In this context, MESs are increasingly required as firms turn to make-to-order, rather than to-stock – with all that means in terms of fulfilling smaller lot sizes, cutting cycle times and developing more flexible manufacturing processes. As Holland says: "Retailers are king now, demanding flexibility and spread of products." And that means both internal change and a need for better 'collaboration' with your supply chain. Indeed, from that point of view, a good MES should be part of the foundation for providing visibility into your supply chains. But there are so many other issues fulfilled by this very large acronym. Suppliers are increasingly mandated to provide more detail on where materials and products came from. And while that's particularly the case in the food, beverage, chemical and pharmaceutical sectors, Malloy sees requirements growing outside FDA-validated environments. "There are other industries where having a very good track-and-trace system is becoming critical," he says. Often such systems are not mandatory, but they provide 'future-proofing' against broadening legislative scope, as well as easing processes around product recalls, service and so on. And MESs don't just aid product traceability, but the effect manufacturing has on the environment. "An MES strengthens your case in proving to regulatory authorities that you are operating within specific guidelines by tracking all constituent parts that go into the product and all actions and events that occurred in manufacture," observes Malloy. Modern MESs also provide everything from two-way communications with business intelligence systems, to overall electronic shop floor management, maintenance management, and optimising all sorts of aspects of automated production in tune with changing demand. In fact, they encompass a hugely sector-specific range of tools. Criteria for success There are eight steps to success. First, plan for cultural change. "An MES system changes the culture and operational parameters of the whole production facility," insists John Bailey, so successful companies have to manage that. Second, vendor/end user partnerships need to be collaborative and strategic, not adversarial and least-cost. Third, standards – use the S95 model, which, as Holland puts it, "best defines the gap between ERP and control in hierarchical terms, and thus what functionality is required." And how to integrate it. Fourth, integrate don't interface: MESs can link the shop floor or process environment with your business/ERP system, but Malloy in particular warns that to get the full benefit they must be properly integrated. "You must have a data map of your system functionality showing what elements of the system need to talk to each other and why," he advises. Fifth, don't go for a 'big bang': Pantek's Richards insists that users want MESs in bite size chunks, and that the tools are there and they should do it that way. Malloy says: "Rockwell's approach is to identify six or seven projects under the MES umbrella that will bring value; then we choose those which yield the highest and most immediate return. It is a modular approach." A key point: you can use the payback from the early projects to fund others, or use it as a test case to convince others in the organisation that the MES you have selected delivers tangible benefit. Pantek's Bailey says: "We go through a prototyping phase – essentially a small implementation with full functionality – which is very good for building confidence. Think of it as a controlled sequence." Sixth, use team-work. "Successful implementations include a cross-functional team steering the project," says Malloy. Best advice: include representation from quality, IT and production to get an all-round view of the MES requirement and its implications. Incidentally, Richards warns that if production hasn't bought into it then it will always be difficult. "They're the guys that are going to use it and who will get the real benefit." There are plenty of examples of companies spending an inordinate amount of time on what amounts to 'banging heads together', and there has to be a better way. Seventh is set targets – so you can measure the project's success. And eighth is take ownership. "Unless there is some really strong ownership from within the end user, it is not going to succeed," says Bailey. "That drive needs to come from the top." Integration, deployment and dealing with legacy systems are, from a functional perspective, relatively easy. Most of the MES vendors believe it's fairly straightforward at all levels, thanks in particular to fieldbus plant networking standards, OPC, the Microsoft platforms and web technologies. As for people and skills, Holland advises: "The up-front collaborative team stage is very much a top-down approach, but the actual deployment/integration work is bottom-up. So you do need to have the core skills to know, for example, how to integrate with older kit on the plant." The trick: establish manufacturing requirements early so that the desired functionality can be identified and integration issues quickly revealed. With Siemens, for example, that task is simplified by using industry templates: integration work becomes more configuration than programming. "Cross-industry libraries address multiple industry issues but can also be tailored for specific sectors," says Holland. And once a library has been developed you can re-use it to deploy standards across multiple, even global, manufacturing operations. That said, Malloy believes integration needn't be a technical problem. "There are some good low-cost integration tools, and provided your architecture is open it can be almost totally vendor-independent." He also says: "An independent system isn't necessarily a problem – you can write an interface or share data fairly easily. If it's older, but still useful to the company, that's a way to go, but you do need to look at the business case to see whether it's more economical to maintain an interface or simply replace that system." Another point: for Richards, the key is deciding what, why and how to integrate to avoid expensive traps. "There's all sorts of expensive middleware and web methods but do you need all that functionality?" he asks. Duncan Fletcher, technical director at Pantek, agrees. "Integration is always a question of compromise. You can have high-speed integration with all the 'bells and whistles' or something that suits the business." Bailey's caveat: integration with legacy systems, such as older SCADA systems, is more of an issue. "You don't want to go into these systems and start changing them just to put in MES." Step forward Wonderware's Industrial Application Server, designed to facilitate exactly that. In any event, the S95 model for MES integration gives a good methodology for mapping out the options for connecting different aspects of your production and business IT. Tools and technologies Returning to the possibilities, typical MES tools include systems for scheduling, audit trails, track and trace, performance management, inventory control, OEE monitoring, and labour productivity measurement and utilisation. Holland outlines Siemens' approach as "laying down the operational procedures, and then drawing from a suite of software that is scaleable," with functionality based on requirement. "The suite is co-ordinated with a modelling tool to allow us to analyse and detail functional requirements using the S95 standard, which allows you to see the interaction between different functions and products." Typically, a client-server approach is adopted; there's also considerable use of Internet technologies for reporting, for example. Siemens' solution utilises Microsoft Visual.Net studio to deploy client applications and a free-formatting HMI to provide data. Beyond that, your key needs are a "good open real-time database that can archive, plus flexible reporting tools to serve information up to various levels within the organisation," says Malloy. "The ability of the database to communicate with the control systems below, and having open architectures, are the keys." Costs of such systems vary from £50,000–500,000. As for ROI, "We've implemented one project recently where payback was six months," reports Malloy, although typically you're looking at more like a year. Final advice? "Focus on getting the maximum value from your IT system," says Bailey. And to do that he suggests: "Look at cost of ownership and how data and software are managed." That way, the disciplines that exist in the IT environment will be transferred the shop floor. Holland reiterates the importance of adopting the standard S95 model first and foremost. "We've had to wait a long time for MES," he points out. "We've not had a strong business argument or any kind of standards to define it, but these are now in place." Time to revisit MES.