KEEP YOUR PACKAGING MACHINES HUMMING ALONG WITH PREVENTIVE MAINTENANCE
By Joseph C. Panettieri
Even if it ain't broke, fit it.
Or so say packaging experts in the know. Much like a car’s need for regular oil changes and timely tune-ups, packaging machinery needs proactive TLC. Harsh chemicals, intense friction and impatient operators (not to mention the humid, dusty, hot or cold environments in which machinery must operate) can push even the best-designed machine to a grinding halt. Rather than waiting for their systems to sputter or completely stall, savvy companies find value in proactive maintenance to keep everything humming along better and longer.
But that’s just the tip of the iceberg. At the Center for Intelligent Maintenance Systems (IMS), researchers are working with industry titans including General Motors Corporation, Intel Corporation, Hitachi Ltd. and Rockwell Automation Inc. to develop predictive maintenance technologies—which forecast and prevent equipment failures before they happen, and seek near-zero downtime. “We must transform from today’s ‘fail and fix’ mode to ‘predict and prevent’ intelligence,” says Jay Lee, Ph.D., the center’s director.
Plenty of companies agree with Lee. More than 40 businesses have partnered with the center, which the University of Wisconsin–Milwaukee, the University of Michigan at Ann Arbor and the National Science Foundation launched in 2000. Surging interest in preventive and predictive maintenance is easily explained: If you maximize a machine’s efficiency, you maximize productivity and potentially minimize the machine’s total cost of ownership, notes Beverly A. Politzer, product manager for B&H Labeling Systems in Ceres, Calif. “And by far, the greatest influence on efficiency is the machine’s reliability,” Politzer says.
Still, predictive and preventive maintenance can be tricky processes. In the age of 24/7 global supply chains, demanding shareholders and fierce competition, few packaging lines can afford to stop for an extended maintenance breather.
On the flip side, if you do not take the time to shut down for maintenance, the costs can be high. Tropicana learned the value of preventive maintenance when the vibration sensors in its new ammonia compressor room sensed a start-up problem with the evaporative condensors, says Glenn Johnson, senior manager of automation and utilities. “The payback on the vibration sensors was almost immediate,” Johnson says. “We are talking a bad bearing on a 1250 horsepower motor. The motor costs over $40,000, and would have probably cost us over $20,000 to repair. In this instance, predictive maintenance measures saved the company a lot of money.”
|
FLAWLESS PERFORMANCE ? FOUR RESEARCH AREAS THAT COULD REVOLUTIONIZE MACHINE RELIABILITY
Source: Center for Intelligent Maintenance Systems |
Getting Started
In order to prevent problems, businesses must first create “baseline” parameters that indicate the machine’s typical performance (say, 700 product labels applied per minute). If actual output declines from the performance baseline (to, say, 680 labels applied per minute), it’s time for a machine operator or technician to inspect the system for problems.
 |
|
Monitoring Technology Corp's 20/20 Hindsight system. |
An alternative would be a fully automated check system. For instance, FactoryWare of Evanston, Ill., offers automatic data collection and analysis software that allows machine operators to rapidly troubleshoot problems. The company’s FactoryPulse software captures, categorizes and analyzes minor stops in a production process, including carton jams and label quality issues.
The good news: Many inspections don’t require system shutdowns because most packaging machines include embedded sensors. Embedded sensors collect data about packaging machines much like a cardiograph, blood pressure monitor and thermometer gather data about a medical patient. As a result, 60 to 70 percent of manufacturing equipment problems can be predicted using operations data and process parameters, asserts Lee.
Tekscan Inc.’s I-Scan system, for instance, includes sensors that measure pressure distribution during a packaging process. The information is fed to and displayed on a Windows-based PC. If the machine is applying too much—or too little—pressure as a product is sealed, the machine operator can make adjustments on the fly.
“By adjusting the sealing forces before the start of a production run, you can avoid large amounts of waste,” says Alison King, a marketing assistant at Tekscan’s industrial division in South Boston, Mass. “The sensors can also inspect for worn parts to help avoid a future failure.”
Other monitoring systems, known as video recording systems, have moved beyond the testing phase and are now widely used by packaging operations managers. Two prime examples include Hartness International Inc.’s HERO and Monitoring Technology Corp.’s 20/20 Hindsight. With video recording systems in place, such problems as setup issues and minor line jams can be corrected before they spiral into larger problems, says Jim Dechman, president and CEO of Monitoring Technology.
|
GOT VISION? The instant replay is moving from the sports world to the packaging world. Monitoring Technology of Fairfax, Va., introduced 20/20 Hindsight in 2002. The system consists of specialized software and a PC running Microsoft Corporation's Windows XP; a flat panel monitor; a CCD (charged-couple device) camera and high-powered light; and a cart to move the system from line to line, which typically costs $21,000 (or $26,000 for a laptop-based version). If that cost sounds prohibitive, consider these two factors: Hindsight claims the system pays for itself within two to three months and early adopters include Coca-Cola, Colgate-Palmolive, General Mills, Kraft and Procter & Gamble. Hindsight takes 60 digital pictures per second, recording each photo separately. Users can quickly find a specific indexed photo using a patent-pending technology known as Scanback. “Hindsight is like having another set of eyes on the floor,” says Jim Dechman, president and CEO of Monitoring Technology. “You just point the camera and walk away from the system. If a line problem occurs, you can quickly scroll through the video to see the jam in slow motion.” “Jams and rejects may be at a minimum on your production lines, and your sensors may be telling you all you think you need to know,” adds Fred C. Erler, executive vice president of packaging at Monitoring Technology. “But a picture of your process could be worth thousands of dollars to you in reclaiming rejected product and improved efficiency.” Meanwhile, HERO International LLC, a subsidiary of Hartness International Inc., introduced HERO (Hartness Error Recording Online) in June 2003. The HERO system consists of interconnected cameras that provide blanket coverage of a packaging system. Prices range from just under $20,000 for a four-camera system with a touch screen to just under $45,000 for a 16-camera system. Rather than working in a vacuum, HERO communicates with packaging machinery PLCs (programmable logic controllers). As a result, the PLC interfaces allow specific camera views to pinpoint errors on the packaging line. “In other words, HERO provides event-triggered video capture,” says Mike Wagner, business development manger for the Global OEM (original equipment manufacturer) team at Rockwell Automation of Milwaukee, Wis. “So it's much easier to pinpoint a problem and make the correction.” Rockwell Automation is offering a private-label version of the HERO system called Flashback, enabled by HERO technology Machine operators aren't the only ones who can benefit from HERO. Managers can use Internet or e-mail connections to review and correct recent production problems and vendors can review video clips remotely before dispatching a service technician. “A system like this can dramatically reduce finger pointing when a problem arises,” says Alison McPheely, vice president of HERO International. With continued advancements, video recording systems will go a long way toward eliminating problems from the get-go. |
“We utilized the system to identify the root cause of jams in a bundler that was plaguing us for weeks, contributing to significant efficiency loss,” says Kristen Kukral, senior reliability engineer, Unilever HPC-NA. “After setting up the system, we were able to identify the root cause of these jams in only 10 minutes. We were amazed at the immediate payback on our investment.”
Unilever uses the Hindsight system to play back downtime incidents and correct any problems. “We’ve been able to move frame-by-frame through high speed bar-soap packaging processes to identify the root cause of equipment stops that we never would have seen with the naked eye,” Kukral says.
Monitoring the production line not only helps packagers correct problems, but also can help identify other areas of opportunity to increase line efficiencies. “Our site was facing significant quality and customer service issues due to production of improperly aligned tray packs for bar soap product,” Kukral says. “Trays were also opening up and jamming on the way to the warehouse, resulting in lost production. The 20/20 Hindsight system allowed us to quickly identify a modification that our supplier could make to the cutout design of our trays to improve the alignment of our tray packs.”
Analyze This
 |
|
Hartness International Inc.'s HERO system |
Unfortunately, staff turnover, poor technician training and plain-old laziness can prevent raw data from becoming useful information. “Most end-users now understand that data exists at just about every point on the [packaging] line,” says Mike Wagner, business development manager for the Global OEM (original equipment manufacturer) team at Rockwell Automation of Milwaukee, Wis. “This has led to vast amounts of data being captured. But in most cases the data hasn’t been used.”
In many cases, companies can gather data from packaging machines and readily share it over the Internet. However, online information exchanges have been limited because many companies prohibit equipment makers from collecting their operational data, says Lee from IMS.
Slowly but surely, that’s changing. Today’s OEMs increasingly partner with suppliers, end-users and research organizations to gather, present and analyze a machine’s data more efficiently. This collective process can predict and prevent problems before they occur. It can also improve the way machines are designed.
For Unilever, the predictive maintenance process helped them with changeover. “We utilized the 20/20 Hindsight system during a trial production run. We wanted to be able to run two different products on the same packaging line without extensive changeover losses. The system helped us identify a component redesign opportunity that reduced the number of change parts for a product change,” Kural says.
Just ask Keene, N.H.–based MARKEM Corporation. Like many manufacturers, MARKEM performs real-world tests within
environmental chambers and at demanding customer sites. The empirical data is analyzed which ultimately helps the company to determine “Mean Time Between Failures” (MTBF) ratings for its systems. “If the MTBF is 17,000 operating hours, then that is truly the MTBF you can expect, assuming you don’t have operator-induced problems,” says John A. Callini, principal engineer.
|
FOUR STEPS TO SUCCESS
|
Penny-Wise and Pound-Foolish
Still, many packaging companies fall prey to well-meaning internal accountants who are preoccupied with operating budgets. In order to minimize operating expenses, some financial pros cut maintenance budgets, veto service contracts or block vision systems purchases.
“Companies often make short-sighted mistakes when they attempt to reduce ongoing operating expenses,” says B&H’s Politzer. “Simple things like failing to purchase spare parts in a proactive manner tempts fate. Sometimes a machine can be down for an extended period of time for want of a very small component.”
Convincing management to loosen the corporate purse strings for preventive maintenance or a several-thousand-dollar vision system is a tricky process. The most effective approach is quantifying the cost of downtime. Present a specific scenario to management, such as a packaging machine failing for 90 minutes. If such a failure occurs, what’s the total financial impact? Be sure to include estimated costs for idle workers, emergency support fees and the value of products that weren’t packaged during the downtime.
“Our industry is being forced to run at higher and higher speeds with fewer and fewer employees,” says Alison McPheely, vice president of HERO International LLC, a subsidiary of Hartness International.
To succeed, that formula requires a substantial bit of predictive and preventive maintenance.
Joseph C. Panettieri is editorial director at the New York Institute of Technology. He has covered Silicon Valley since 1992.
