Knauf Insulation is a German-based company with a fiberglass manufacturing plant in Shelbyville, Ind. The company has recently connected part of its plant data system to a Microsoft SQL Server to monitor product quality. The plant produces thermal and acoustic fiberglass insulation for the residential and commercial markets. Products include fiberglass pipe insulation, insulation board, duct wrap and tank insulation. One of the critical measurements the plant needs as a way to assure quality is the density of fiberglass in the product. “We have very rigorous quality standards,” declares Curtis Davies, a corporate process engineer at Knauf. “Any deviations from in-house materials specifications on any of our products can result in scrapping the products, which is a waste of many man hours.”
Plant operators use a nuclear-based gauging system for density measurements. The business side of the operation wanted to import the data from the gauging system into the SQL Server so managers could build product profiles for analysis and optimization. Operators also wanted the data so they could make on-the-fly adjustments to the manufacturing process. Not surprisingly, complex engineering would have been required to enable the gauging system to communicate the data to the SQL Server. Plant engineers switched the plant’s control system hardware to input/output (I/O) devices from Opto 22, Temucula, Calif., because the Opto 22 controllers use open, standard protocols.
The company also uses software from Wonderware, a Lake Forest, Calif., vendor, which was used to interface with the Opto 22 devices. The Opto 22 system was able to network with other intelligent controllers—
regardless of their original manufacturer. The result was networked connectivity from the control devices to the SQL Server. In all, Knauf is now able to monitor and control more than 30 I/O points, which offers up data to the SQL Server for analysis.
Simple it’s not
As plants begin connecting their manufacturing data to other company information systems, the Knauf experience is typical. Plants exchange data with the business side of the enterprise to solve one problem at the time. Connectivity from control devices to other information systems such as SQL Servers does not necessarily come easily, and the rationale for sharing plant data is usually done on an item-by-item, need-by-need basis. It’s not as simple as plugging the automation system into the enterprise resource planning (ERP) system. Usually, plant operators have to be dragged to the meetings where business people ask for plant data. “What we’re seeing is plants dealing with individual points of pain,” says Ron Monday, chief executive officer of Online Development Inc., an automation connectivity products vendor in Knoxville, Tenn. “The connections are usually driven by the front office, and grudgingly, the shop floor complies.”
Monday notes that different industries have different reasons to connect plant data to the enterprise. Some want to improve efficiency and shave a percentage off manufacturing expenses to improve the bottom line, while other industries make the move in order to comply with regulations. “They buy the connectivity to solve points of pain, or to take the perceived enjoyments of reduced costs,” says Monday.
Jay Coughlin, product manager of HMI at Siemens Energy and Automation Inc., the big Alpharetta, Ga.-based automation vendor, agrees that each individual industry seeks plant data to solve different problems. “The automotive industry is looking for a little increase in efficiency,” says Coughlin. “That’s different from the textile industry, which has a lot of manual processes and doesn’t need as much information. Food and beverage, and chemical—they’re big on asset management, while the pharmaceutical industry has to meet regulation oversight.”
Serving customers more effectively is also a significant driver that’s sending business people down to the shop floor for data. “Companies have to be able to tell the customer whether they can make the product when the customer needs it,” says Colin Masson, research director of manufacturing operations at AMR Research Inc., in Boston. “The business people need to know if the machines are fully loaded, whether they have finished goods, whether they have the raw materials and production capability. If you want to know that in real time, you need automated data collection.”
Over the past 10 to 15 years, companies have used their ERP systems to connect the enterprise from one end to the other. The first connections went naturally to the office functions: financials, human resources, sales and marketing, supply chain and logistics. The final mile of connectivity goes to the plant itself. The information technology (IT) department knows, of course, that the plant runs on an information platform, but until recently, plant data has been its own animal.
As companies struggle to become more competitive, they are now looking to see if their plants can become leaner and more adept at meeting customer demand. “They want to be able to respond to changing demand and become a more proactive organization,” says Marc Leroux, manager of collaborative management at automation vendor ABB, in Warminster, Pa. “They’re looking at the plant data and they say, ‘We’re making real-time decisions on month-old information.’
“So they turn to the plant operators and say, ‘Let’s get this connected to the ERP system.’ ” But the plant automation system and the ERP system are not so easily connected. “ERP systems made it quite easy to connect to other ERP systems, but not to the factory floor,” says Leroux. “[ERP vendors] Oracle and SAP knew the plant data was available, but they figured they could load it up every month or so.”
Part of the reason plants have been slow to connect their shop floor data to their business systems is simply because the IT department had its hands full in recent years just connecting the business functions to the ERP system. “Connecting to the plant usually comes later in an ERP implementation. It ends up as an add-on to the ERP," says Kevin Tock, vice president of production and performance management at Wonderware. “Sometimes it takes two years.”
Business needs data
Yet the business side now realizes it needs data from the plant to monitor and trim direct costs. “You can use plant data for real-time performance management,” says Kevin Zamzow, product manager for wireless technology at ProSoft Technology Inc., in Bakersfield, Calif., a wireless solutions company. “If you have the data, you can determine if something has happened at the plant, you can see what’s causing variation and adjust for higher quality products and save energy. You can also monitor recipes. A steel company was able to record information on a recipe for steel, and by automating the plant to hold to the recipe, they were able to eliminate errors.”
Of course, the notion of the business side monitoring and adjusting plant operations to optimize performance strikes terror into the heart of plant operators. In many cases, perhaps most cases, the IT department and plant operators have not been eager to work together to develop plant-to-business connectivity. Mutual mistrust between plant management and IT executives is common across most industries.
One reason is that the plant has mostly been left alone over the years to develop its unique automation technology. That situation changes when the IT department peeks into the plant in search of data. IT is nervous about security as the plant adopts Web-based connectivity, and the plant operators are worried about safety as IT pokes around in plant data. “IT keeps talking about going after data,” says Monday of Online Development. “And plant operators say, ‘Stop. One of these days you’ll be going after data and you’ll cut someone’s hand off.’ ”
Some believe that IT will win the battle, as long as the IT department is well funded and savvy about connectivity. But in order for this awkward marriage to work, the IT department has to convince plant operators that its use of plant data is not going to directly affect plant operations. “In the future, the IT departments with more money and higher connectivity within the enterprise will take more responsibility for the direct connection with the factory,” says Trayton Jay, director of special projects at vendor Mitsubishi Electric Automation Inc., in Vernon Hills, Ill. “The manufacturing people say, ‘I don’t want them to have the raw data.’ But the IT people say, ‘It doesn’t matter. The data is indirect.’ ”
Standards are the key to the connection between a plant automation system and business functionality. But standards alone—and connectivity alone—do not justify the awkward trading of data between the plant and front office. The connectivity decisions are being made and justified point-by-point, problem solved-by-problem solved. The front office needs specific data to comply with track-and-trace regulations. Business needs win the justification for connectivity. Newly emerging standards provide the connectivity.
For more information, search keywords “shop floor-to-top floor” at www.automationworld.com.
To see the accompanying sidebar to this story - "OPC or Database: How are People Connecting?" - please visit www.automationworld.com/view-3000