Fortunately, the situation is changing, led by articulated-arm robots for case packing, palletizing, depalletizing and other material-handling applications. The shift is due in part to a manufacturing environment where menial jobs are hard to fill and lean manufacturing encourages the use of machines for non-value added jobs. Another factor is the maturation of robotics technology and advances in complementary technologies. According to a recent survey by the Packaging Machinery Manufacturers Institute, 35 percent of manufacturers plan to purchase robots this year, with palletizing robots the most likely application.
"It wasn't a matter of getting rid of the labor; the labor was disappearing," Fallin adds. "The repetitiveness of some of the work was an ergonomic nightmare."
Two dozen potential initiatives were identified in Fallin's first six months on the job. Robotic loin trimmers and a self-sharpening blade using technology from Sandia National Laboratories were rejected as too pricey or impractical for deployment. An automated squeezer of pig intestines produced chitlins too clean for customer expectations and was scrapped. The project eventually focused on the packaging area as a prime candidate for robotics.
In 1997, Fallin began working with Andrew Courier, a design manager with Foster-Miller Inc., on a robot capable of packing a wide variety of primal cuts into shipping cases. The cuts arrived in a pack-off area in vacuum bags and had to be oriented to customer specifications, typically with a target weight of 100 lbs. per box. End effectors such as suction cups and pneumatic claws were defeated by the cuts' random shapes and weights and by the package material. After studying the way operators packed, the engineering team developed a bucket with four axes of movement and an air bladder to secure each piece. "The (bucket) elevator goes up and down and tilts for a sixth degree of freedom," says Courier, a mechanical engineer. "The algorithms for queuing, sorting and orienting each piece resulted in a fare amount of autonomous intelligence."
While observing the workers, the team noticed they relied on touch rather than sight, which inspired engineers to install sensors to orient and handle the cuts. "Vision systems have their place," says Fallin, "but they don't have the cognitive ability of the human brain. A vision system can't generalize."
"Vision systems add a level of complexity, there's no doubt about it," agrees Tim Schiller, key technology manager-robogistics at Clinton Township, Michigan-based Kuka Robotics Corp. "But vision adds a lot of power and flexibility, and unless you're getting involved in meat cutting, vision systems work well. And calibrating a camera for a robot is getting easier and easier."
Digital cameras compute pixels, while a robot relates to the world in millimeters. Translating those dimensions required significant engineering in years past; now plug-and-play solutions are surfacing. Additionally, some cameras can be cabled directly to a robot, eliminating the need for a separate processor for the camera. And today's processing power is magnitudes greater. "In the old days, you ran out of (computing) room if the robot had to process more than 500 kilobytes," says Schiller.
"Machine vision using stills and surveillance technology using moving images are coming together to allow manufacturers to track items coming down a (conveyor) line," says Dave Holland, a production manager at Fonterra Cooperative Group Ltd., a New Zealand dairy that deploys 1,283 PLCs in its facilities and has a controls infrastructure valued at $320 million. Fonterra has started using cameras with articulated-arm robots for depalletizing. Digital cameras also are being incorporated into automated guided vehicles (AGVs) to store and retrieve inventoried product.
Fonterra abandoned gantry systems 10 years ago and embraced articulated arm units because of their lower capitalization costs, pattern-forming flexibility and reprogramming advantages, Holland said in a presentation at last year's Worldwide Food Expo in Chicago. The company now has 60 such robots, including a two-robot, 10-lane system recently installed to palletize 55-lb. cases of butter. "It's a complex system that replaces 10 people and handles up to 1,200 boxes an hour," he said. "It's also rugged: the parts that tend to break down are the conveyors and box handling equipment, not the robots themselves."
QComp integrates ABB robots with a Danaher controls platform used "in tens of thousands of AGVs worldwide," Doyle says. Logic embedded in the program simplifies reprogramming of the AGV's path. Laser-guided systems offer the greatest flexibility. The lasers send a signal every 50 milliseconds to wall-mounted reflectors to triangulate their position and stay within one-quarter inch of trajectory. Speeds of up to 3.5 ft. per second are attained.
While OEMs emphasize the flexibility of articulated arms, they acknowledge the limiting nature of end-of-arm effectors. "That is the make or break of a palletizing robot," notes Laxmi "LP" Musunur, engineering manager of palletizing and packing at Fanuc Robotics America, Rochester Hills, MI. "That is where a good systems integrator comes in."
About 80 percent of robotic palletizing involves corrugated cases. Vacuum effectors are great for corrugated, but club stores and other retailers are demanding display-ready cases. "There is a move away from vacuum technology because so much food and beverage product is shipped in trays and shrink bundles today," points out Pat O'Connor, product manager-palletizing systems at St. Louis's FKI Logistex NA. Fork-and-clamp tools provide more positive grip and lift, though the trade-off is slower speed.
One solution is a hybrid effector, which achieves 20 cycles per minute, Musunur suggests. Mechanical assists counteract the lateral shear from rapid acceleration and deceleration that can defeat a suction system.
Incremental changes like freezer-specific robots and carbon-fiber forearms for longer extensions without excessive arm weight will pave the way for more food applications, says Probst. Familiarity with palletizing robots also will increase use. He likens robots to servo motors. In 1976, Probst installed the first servo for a computerized bologna slicer in Oscar Mayer's Davenport, IA, plant. "That was scary, state-of-the art technology at the time," he recalls. "Now it's commonplace." The same maturation is underway with robots, with articulated arm palletizers playing the role of servo-driven bologna slicers.
For more information:
Rick Goode, Columbia Okura LLC,
Laxmi "LP" Musunur, Fanuc Robotics Corp.,
Pat O'Connor, FKI Logistex,
Andrew Courier, Foster-Miller Inc.,
Tim Schiller, Kuka Robotics Corp.,
Tom Doyle, Qcomp Technologies Inc.,
Paul Probst, von Gal Co.,