In manufacturing’s digital age, data availability never is an issue. Information availability, on the other hand, is another matter.
The challenge goes beyond polling data from PLCs, flowmeters and other process control devices. Records of process variables need to be analyzed in the context of outcomes. Marrying lab data to process data requires either a legion of key-entry personnel or an investment in business and process software. Manufacturers may not know what they don’t know until a capital investment is made, and that can require a leap of faith.
Milk Specialties Global took that leap last year. “We’re lean in number of people,” says COO Brian Lundquist in describing the 68-year-old dairy protein company, which is based in Eden Prairie, Minn. (www.milkspecialties.com). Long active in animal nutrition, the company made a well-timed move into whey and milk protein concentrates nine years ago, helping to push annual sales volume near $1 billion. Growth in protein-concentrate demand is expected to “be double digits for the near future,” according to Lundquist.
When order fulfillment is a priority, efficiency and yield tend to suffer. Milk Specialties’ two-person process excellence department spent much of its time collecting log sheets and keying numbers into spreadsheets for analysis. Process adjustments took time, and raw material variability often undid the assumptions on which they were based.
To improve outcomes, Milk Specialties installed a data historian and business analytics software at three of its six plants. The goal was to pair data from the laboratory information management systems with real-time production data. Of particular interest was performance of filtration systems and the best times to either clean or replace membranes.
Membranes are expensive, and running a clean-in-place cycle necessarily involves downtime. Scheduling the optimum time to do either has a direct bearing on profit margin, points out Jean-Luc Bonnet, a regional manager with Milwaukee-based Rockwell Automation (www.rockwellautomation.com), who participated in the project. "Manufacturers know more or less what they have to do, but they waste time and money if they don’t have real-time information on when to do it,” he says. Analytics provided the key performance indicators to guide what actions to take and when to take them.
Better filtration management contributed to the $750,000 a month in cost reductions that Lundquist says Milk Specialties realized. Another contributor was improved yield, with more of the available protein in milk and whey finding its way into finished goods.
One of the three plants affected is in Visalia, Calif., where industry faces stiff fines and surcharges for exceeding limits on chemical oxygen demand and total dissolved solids in wastewater streams. Milk Specialties’ facilities now reuse much of the water removed from milk and whey, helping cut municipal water consumption 40 percent. Between reduced consumption and lower discharge fees, the firm reduced costs $100,000 per week.
Besides giving plant personnel actionable information, the system gives the process excellence department insight into the relative performance at those locations. As a result, best practices are identified and transplanted to the other sites.
Business intelligence software is a tool, Lundquist summarizes, but it’s only as good as the people who use it. “BI points you to the lake where you should fish,” he observes, “but you still have to catch the fish yourself.”
Indiana’s newest mogul
Companies engaged in buying and selling seldom segue into making, but that’s precisely the direction Queen City Candy Inc. started moving to in 2000, when the candy broker became a rebagger. Copacking capability was just a warm-up to 2015’s transition to manufacturing of gummies, jellies, jujus and fruit snacks.
In 1983, president Vince Klee founded the Greendale, Ind., firm, located 15 miles west of Cincinnati. Retailers don’t want to manage relationships with multiple confectioners, so brokers like Queen City buy in bulk and consolidate orders and shipments. Klee expanded into copacking with a climate-controlled warehouse and the purchase of a vertical form/fill/seal machine.
Referring to Queen City’s entry into manufacturing, Klee says, “It’s been a goal for the last 10 years, but it didn’t get serious until four years ago when our gummy manufacturer had an explosion that killed eight people.” The Oct. 24, 2013, explosion at Dulces Blueberry plant in Ciudad Juarez, Mexico, was attributed to a chemical reaction involving corn starch.
Gummies production was targeted in part because “it’s a really user-friendly media for nutraceuticals,” Klee explains. A Queen City customer now exports gummies with vitamin supplements to 20 countries, using a formulation with sufficient heat stability to withstand ocean transport.
Pharmaceutical standards guided 2015’s construction of a 10,500-sq.-ft. processing plant to house a kitchen, pressurized starch molder (a mogul, in confectionery parlance) curing room and other production systems. Queen City’s mogul has a capacity of 10,000 lbs. per hour, about 1 million gummies.
Though not the highest volume mogul from Robert Bosch GmbH, Queen City’s unit will index 35 trays a minute and is better suited to the frequent changeovers required by a contract manufacturer, notes Robert Dono, national sales manager-confectionery for Bosch (www.bosch.com).
Production started in October 2015. Anthony Habib, a consultant who serves as Queen City’s process engineer, oversaw the installation and systems integration work, using a distributed controls system from Siemens Inc. Workers now use iPads to record quality data from 18 process-control points.
Greenfield confection projects are a rarity in the U.S., where high sugar costs have driven scores of manufacturers to Canada, Mexico and other countries. As a consequence, packaging that proclaims “made in USA” has acquired special cachet with candy buyers.
A distributed control system from Siemens Inc. (www.siemens.com/food-beverage) in Atlanta provides machine data and feedback from sensors that monitor cook temperatures, pH readings, viscosity and curing conditions. When a steam trap in the kitchen became blocked, reducing thermal capacity, software helped pinpoint the source of the problem before quality issues occurred, reports Habib.
Once production started, coding work on a customized enterprise resource planning (ERP) system got under way, replicating the path taken when a warehouse management system was installed to track inventory. “I was skeptical” about deploying customized ERP, Klee admits, but at the recommendation of a business associate, he commissioned Lalit Agrawal, president and chief software architect at Cherry Group Inc. (www.cherrygroupinc.com) in Mason, Ohio, for the project. Instead of modifying procedures to fit an off-the-shelf ERP package, the custom program bends to the company’s methodologies and delivers information in the favored format.
Elements of an ERP system already were in place, but integrating machine data into a business information system with sophisticated tracking capabilities was new for Queen City. “It’s a completely customized system,” adds Agrawal. “We do not have to play by the rules of the software. There’s no dependency on a black box.”
Brand owners demand co-manufacturers adhere to the highest food safety standards. Queen City management sought SQF Level II certification, an endorsement that would have been extremely difficult to attain if automated data collection and reporting were not in place, Habib suggests. Recall capabilities are a crucial part of the SQF audit, and manufacturers are given a two-hour window to trace a given ingredient back to the supplier and forward to the products in which it was used.
“Our customers wanted us to prove that we could do what we said we could,” adds Klee. “Now, everybody is knocking on our door.”
Bullet-proof recall systems
Traceability is part of food safety, and customer expectations are high. A team from one prospective customer spent two days recently at Queen City, auditing the plant and reviewing the tracking system.
A robust recall system can turn an event into “something annoying instead of catastrophic,” points out Doug Renfro, president of Renfro Foods Inc. (www.renfrofoods.com), a salsa and sauces manufacturer in Fort Worth, Texas. He speaks from experience: Two recalls triggered by ingredient suppliers and affecting a customer’s order became non-events thanks to the information system in place.
“The customer was impressed by the data we had and how quickly we responded,” Renfro recalls. What the system lacks in glitz, it makes up in functionality.
Renfro’s system was installed a dozen years ago by Tamlin Software Inc., a Dallas firm catering to small and mid-sized food manufacturers. Traceability is one module in the process control program, which also captures quality data such as hot-fill temperature and vacuum, pH level, viscosity and lab test results. “A baby version” of statistical process control is embedded, says Renfro.
While the software vendor provides upgrades every two years and customized reporting capabilities when requested, Renfro characterizes the system as mainstream, not leading edge. There are no tony LAVISH user conferences in posh venues to share experiences, only informal meetings to swap stories with other food companies using the same software.
Nonetheless, the system meets the record-keeping requirements of the Food Safety Modernization Act and the demands of SQF auditors. Half of Renfro’s business involves copacking. The information generated satisfies those customers, leading Renfro to conclude, “If it ain’t broke, don’t fix it.”
Data fundamentals haven’t changed in decades, allows Walter Staehle, national and strategic account management at Siemens and a former director of manufacturing systems at Kraft Foods. What has changed is availability: who has access to the information generated and who is empowered to act upon it.
In the past, workers collected data on the plant floor and passed it upstream, where decision makers mulled trends and passed down directives, Staehle remembers. Today, plant workers “can make instantaneous decisions” to correct process deviations. “The people that own the process are the people who are running it day after day,” he says. Electronic collection makes data more reliable, but changes in work practices make it more meaningful.
Company size, complexity and needs dictate the best software fit for a manufacturer. As integration of programs for business management and process control grows closer, the value of information derived from all the data is growing.