Engineer-Centered Supply Chain
Wait a minute. Wasn't all this shop-floor-to-top-floor talk put to rest in the late 1990s when many manufacturers, large and small, could not find a way to make this work without spending millions of dollars and enduring years of integration work? And isn't it true that most companies never got the return they expected from their enterprise systems, with many not even turning on most modules in...
This article contains online extras.
Wait a minute. Wasn’t all this shop-floor-to-top-floor talk put to rest in the late 1990s when many manufacturers, large and small, could not find a way to make this work without spending millions of dollars and enduring years of integration work? And isn’t it true that most companies never got the return they expected from their enterprise systems, with many not even turning on most modules in the systems they purchased?
Yes, the brutal realities of plant-floor-to-front-office integration did put a damper on the enterprise-wide connectivity hype. But the urge to integrate never really went away, and here’s why: As the economic bubble burst and manufacturers retrenched operations in reaction to one of the worst industry downturns in recent memory, a number of companies discovered that the ultimate effectiveness of their supply chain depends on real-time information from the plant floor (see ‘Real-time’ clarification sidebar).
Despite this realization, the problem remains that supply-chain systems are transaction-based—dealing with operations and production information on a weekly, daily, or possibly an hourly basis. To respond to the disappearing practice of carrying large volumes of inventory and the accompanying increase in related customer demands, manufacturers have to synch their supply-chain operations with real-time shop-floor realities. Otherwise, they can count on losing business to more flexible and agile competitors. And though relationships still go a long way in sustaining a healthy business operation, loyalty in business largely disappeared with the arrival of the on-demand world. To succeed today, your company has to deliver what the customer wants, when they want it, how they want it. That can’t be done without knowing exactly what you’re capable of delivering; and that’s why the focus is now squarely on the shop floor and the engineers who run it.
‘The whole idea of the demand-driven supply network is that you and your supply chain partners have a network that’s striving to take advantage of opportunities out there to get the highest margin you can,’ says Bill Swanton, vice president of research at AMR Research. ‘If you’re going to do this in a global world, you have to understand what the capabilities of your manufacturing operations are. You need to have visibility into what capacity is available or what the trade-offs to using those capacities in different ways are. You need to understand what yields you can expect and just what you could do. And all of that comes down to having visibility across multiple plants. Not only how they’re performing, but what they’re doing and could be doing.’
In this speedier, on-demand environment, companies’ attention and project spans are getting shorter, thereby driving focus directly onto processes that immediately affect outcome. ‘Manufacturers want to be cognizant of exactly what is going on in things like WIP (work in process) and capacity utilization,’ says Yves Dufort, general manager, Wonderware Strategic Integration Group of Invensys. ‘For example, [manufacturers] want to peg their manufacturing order from the ERP system and know what state it is in at all times and what capacity areas it is utilizing at any given time. That’s all tied to market responsiveness to be able to determine if a product is gaining traction in the market.’
Matt Bauer, Interwave director of business development and marketing, Rockwell Automation, says that more and more Fortune 500 companies are approaching how to deal with an on-demand information environment strategically. (Interwave is Rockwell Automation’s manufacturing execution system [MES] software.) ‘That’s a big change,’ he says. ‘It’s become more strategic now rather than a tactical answer to a specific capacity or quality problem, because it’s not just on-demand driving more access into what’s happening on the shop floor, it’s also ‘lean initiatives’ and ‘Six Sigma.’ ‘
Though most people are aware of the supply chain benchmarks set by Dell Computer, many don’t realize how ubiquitous the drive toward lean supply chain operations has become. ‘Even companies like Eastman Chemical are doing this,’ says Dave Smith, director of strategic marketing for OSIsoft. ‘They are producing a lot of their chemicals to order rather than to stock—especially the specialty chemicals, because they are so expensive and so are the raw materials that go into them. Standing inventories are just not acceptable anymore.’
Regulations are also having an impact—and not just in the consumables market where U.S. Food and Drug Administration’s 21 CFR Part 11 (on electronic records and signatures) is having far-reaching effects into production processes. The National Highway Traffic Safety Administration’s TREAD (Transportation Recall Enhancement, Accountability, and Documentation) Act requires automotive suppliers whose products have any impact on consumer safety to track and report any safety-related issues, such as the number of warranty requests and consumer complaints. Once these data are reported and the NHTSA looks through them, if regulators decide to investigate further, they’ll ask to see product genealogy in terms of what’s causing the safety issue to occur.
‘Many customers have manual production lines and are therefore not recording product genealogy info,’ says Mark Doyle, vice president of I/Gear (a provider of hardware and software for collecting, integrating and utilizing plant-floor data). ‘For example, if you’re making a car seat, are you recording what batch of seat-belt products you are using to go with that seat assembly, so that if you do need a recall, you can quarantine those seats in the smallest amount based on the product or part you determine to be causing the issue?’ he asks. ‘This is the kind of stuff that’s causing manufacturers to evaluate how they collect transfer and access data.’
Consolidation in the manufacturing industry is creating additional demand for plant-floor data. ‘To manage a consolidated environment with reduced workforces, you’re asking fewer people to do more—so you have to have a good automation strategy and visibility of production info to make good decisions about what actions to take,’ says Kevin Roach, vice president of Global Solutions, GE Fanuc. ‘Also there’s the issue of asset utilization. A lot of organizations, through consolidation, are recognizing they are over capacity. So how do you recognize where you can most effectively improve existing assets and potentially reduce or remove underachieving assets ?’
How it’s being done
The ‘how’ part of enterprise-wide linkage really began with engineers working to accommodate requests from the IT and corporate sectors.
As more vendors offer packages designed to move data between the plant floor and enterprise systems, the analyst groups have provided names to classify these types of software applications. ARC Advisory Group calls it Real-time Performance Management (RPM). RPM is the process of providing corporate officials with top-level performance measures through real-time accounting systems and operations data provided from the supply chain and associated plant-floor activities.
AMR Research refers to this as Enterprise Manufacturing Intelligence (EMI), which defines the role of software in this space as providing five basic functions: 1) aggregating plant-floor data; 2) contextualizing those data; 3) analyzing data; 4) propagating data; and 5) permitting visualization by drilling down as far as needed. (See ‘Basic EMI capabilities’ graphic).
In the end, this class of software is all about getting a handle on plant-floor data so that better business decisions can be made, linking the engineering and business sides of manufacturing.
‘EMI vendors are focusing more on visibility, data connections, and taking data from multiple sources,’ says Swanton. ‘The nice thing about EMI, unlike MES, is that it does not change the way you process orders in your facility, so you’re not taking a risk of interfering with production. This appeals to engineers and management alike.’ Risk is limited to installation cost and benefits typically pay back the cost in fewer than six months, Swanton says. (See ‘Return on EMI investment’ graphic.)
To understand how EMI can benefit engineers, consider that one of its major functions is overall equipment effectiveness. EMI can provide a statistical basis for analyzing performance of all plant assets. And if your plant is using an enterprise asset management (EAM) system that has the capability of triggering maintenance activities based on use, EMI provides a way of automating that data collection.
Return on EMI investment
|Payback potential (per year)||EMI Project Scope||Description|
|Source: Control Engineering with data from AMR Research|
|10x||Manage fleet of plants or assets||EMI application provides globally distributed organization with standard interface to/and normalized view of performance across disparate existing information systems and architectures. Captures performance information in context of variable costs, market conditions, margin fluctuations, and asset performance. Provides appropriate levels of intelligence and decision support for the various organizational stakeholders.|
|5x||Manage plantwide performance||EMI application aligns business systems with real-time events in production operations, provides decision support tools for procurement, shipping, and inventory management. Improves organization’s ability to respond to and service customers.|
|1x||Tactical OEE||EMI application focuses on visibility into details of line and machine performance with emphasis on overall equipment effectiveness. Identifies opportunities to streamline operations, heighten efficiency, and improve throughput.|
Impact on engineers
Having the spotlight turned squarely on plant-floor data is both good news and, potentially, bad news. The fact that manufacturing engineers hold the key to the next level of success for the business has definitely altered the landscape of players and decision makers.
‘The CIO has never been exposed to the granularity of plant information processing,’ says Jonathan Kall, Interwave director of Information Solutions Americas, Rockwell Automation. ‘And they are hungry to learn more about MES and EMI systems because ERP systems did not return the shareholder value they promised.’
As result, the people in the plant—all the way down to the control engineer—have much more influence on the process than they did in the past.
Being in the spotlight, however, necessitates a slightly less traditional mindset for manufacturing engineers. A wider view of operations is now required, as opposed to being more focused on managing and controlling plant-floor operations and data.
‘They have to wonder about what is a supply chain, what is supply chain integration, what is agility,’ says Roach. He suggests thinking about this new area of EMI software as the next logical evolution after replacing relays with PLCs and dumb terminals with HMI/SCADA.
‘Control engineers understand this and many have lived it. Now they have to transform that story to a different set of objects and those objects are multiple factories, supplier and customer warehouses, customer’s manufacturing lines, raw materials suppliers and transportation systems,’ says Roach.
Roach says engineers shouldn’t feel threatened by this top-floor-to-shop-floor linkage. They should be excited because they have the real-time experience and this is a great opportunity for engineers to expand well beyond their traditional realm of influence. This is new ground for the IT players. As enterprises strive to become more real-time, engineers are the gatekeepers to the history, the pitfalls, and success stories about what real time means.
Another advocate of the upward potential this area holds for engineers is Kall, who has personally seen engineers advance their careers through integral involvement in this process. With one client in particular—a paper converting manufacturer—Kall says that a plant engineer took the time to educate upper management levels about how the company’s SCADA and HMI systems could be used to automatically collect plant-floor data and propagate data upward to feed the MRP (material requirements planning) system with consumption information to make MRP scheduling more reactive to the plant’s real-time reality. This engineer’s idea led to a 10-plant implementation, and he ended up getting promoted several times. ‘He’s now with the company’s corporate global supply-chain operation,’ says Kall. ‘This can be a huge growth path for engineers.’
Real-time: a point of clarification
When the term ‘real-time’ is used in the context of enterprise-wide connectivity, the real-time world measured in milliseconds that some control engineers deal with is not exactly what’s implied. However, it does refer to the ability to drill down into or poll plant-floor data every few minutes—or as often as desired by the user. While that may not be real-time in the engineering world, it is real-time to front-office operations that have historically accessed such data, typically, on a daily or weekly basis.
What’s in the way? Politics of data ownership
Any change in business is often painful to some degree. The push for plant-wide visibility is no exception.
There are two issues creating problems for visibility proponents, according to AMR’s vice president of research Bill Swanton. First, you have to get agreement on common performance metrics—how to measure them and what information is required. Once this is done, you can see how each similar group is doing around the world. Second are change management issues. ‘If you want it [visibility initiative] to be successful, you can’t have the VP of operations in Chicago calling someone in Europe when he sees a line go down,’ Swanton says. ‘It’s not OK to be on their case for every alarm that occurs.’
Yet another issue is: Who owns the data? Is it IT or engineering?
Answers to these questions vary so much simply because companies differ so greatly. ‘If it’s a very ERP-centric organization, the engineer may feel that the only data he has access to that are of value is the process information in the historian,’ says Invensys/Wonderware’s Yves Dufort. ‘If it’s the other way around, maybe the IT guy will feel threatened to a certain degree and not collaborate as much. We’re doing a project now at a pharmaceutical company where engineering and IT are arguing over disaster recovery backups. We try not to make it a battle. If you agree on the metric—you can each have your own data, and it’s not an issue.’
At the root of it all is one simple fact: there is just too much information available today for most people to manage. ‘So how do you make sense of it and deal with data concordance [understanding the information in context]?’ asks Michael Leroux, principal consultant focusing on collaborative production management, ABB. ‘I worked with a customer two years ago on a project where we were taking information from seven manufacturing facilities, integrating them, and providing visibility across the corporation. One of the first things we encountered was something the company called ‘production.’ It turns out that the definition of what ‘production’ meant across the plants was different. So understanding the context, and not just accepting it as a definition understood by all poses a problem—and that’s where usage of such standards like ISA S88 and S95 [for batch and recipe management and enterprise-control system integration, respectively] can provide a common understanding.’
Visibility startup targets the mid-market
If you needed further proof that the whole enterprise-wide visibility initiative is gaining steam, consider that a word not used much these days—startup—is again being heard in the visibility market. The difference with many newer companies in this area is their market focus on the small to mid-range companies. Most of the larger companies targeting the visibility market remain focused on larger players.
One such company targeting the mid-market is Slipstream, launched by Gary Hopkins, a veteran of Citect and Rockwell Automation (where he helped launch DeviceNet).
“If you look at mid-sized manufacturers today, they have a real problem,” says Hopkins. “They have to make improvements to their process, and they’ve done as much as they can in a lot of areas, and now they need to start automating parts of the floor. But to get wider data integration started, they’re typically facing a $100,000-plus implementation. I think that’s why MES has had such a hard time. The Fortune 500s could do it, but it never caught on with the mid-market.”
Hopkins says that analysts only see this space getting more complex, thereby keeping most of the market out of it. But mid-market customers are saying they need to do real-time reporting and work on their KPIs (key performance indicators), but they don’t have the money to start doing it.
“Slipstream allows a single seat to be set up for $7,000 in the first year and provides real-time reports and real-time analytics—allowing the company to get its toe in the water, and as it gains that knowledge and is ready to expand into root cause analysis or track and trace, it just adds [functionality],” says Hopkins.”
Another factor stressing the importance of this level of data sharing for smaller companies is the elimination of old engineering groups. “The number of people holding expertise of what will truly make a line more productive is becoming smaller and smaller,” says Hopkins. “If senior management understands that, it’s going to look more at the engineering side to determine what will be the best implementation piece. There’s been too much enterprise and not enough plant-floor focus. At the end of the day, what matters up top is having the analytics to do proper costing, and it’s the engineers that have and understand the input required for these analytics to be accurate, not the IT guy.”