When Dave Watson first delved into robotics in the late 1980s, these clunky, overly expensive systems were better designed for building cars than making and packaging sandwich cookies at one of the nation’s top bakeries where he worked at the time.
Today, however, these increasingly nimble and often more affordable systems can deftly pick, place and package with precision, but bakeries need to do their homework to make sure they have the support to maintain them and ensure they’re the best investments for their operations.
“There have been tremendous strides on many different fronts from end of arm technology (EOAT) to vision systems to the hardware and software,” said Mr. Watson, food, bakery and snacks engineering subject matter expert for The Austin Co. “All have improved greatly over the past 15 to 20 years or more.”
Specifically, soft robotics pneumatic grippers are designed to pick and lift products based on a certain amount of grip pressure, just like a human hand, noted Jerry Buckley, south central sales manager, BluePrint Automation.
“These end effector grippers allow for a variance in product sizes that are much greater than typical mechanical grippers used in the past,” he explained. “Second, the use of 3D vision camera systems that have become much less expensive allows for robotic systems to better determine if products are within the proper specs.”
Mr. Watson said robotic packaging systems also rely on AI technology to adapt to their surroundings.
“If you look at vision systems over the past five years, they’ve gone from very basic vision to 3D and color systems that allow you a greater ability to identify bad products and even detect quality issues like the color of products as well as to better locate products and pick and place them into packages,” Mr. Watson explained.
By gathering valuable information on a product’s position, color, shape and topography, the vision system trains robots to “learn” how to handle a variety of baked goods, suggested Sushant Sharma, sales application engineer, ABI Ltd.
“Our vision capabilities also add the ability to inspect product as it is packaged. Any defective goods are identified and not processed,” he said.
Yet another advancement involves enhanced control platforms to run the entire system.
“Instead of having two separate machines, the industrial PC has been eliminated, and everything is run through the PLC,” said Bill Kehrli, vice president of sales and marketing, Cavanna Packaging USA. “The footprint for controllers is greatly reduced, and with fewer controls, they’re easier to debug. You don’t need two experts where one knows the robotic controller and the other knows the PLC. The two now communicate great together.”
Mr. Kehrli cautioned robotics are not for everyone and can get pricey on some high-speed lines. For example, a granola bar line that runs 1,200 pieces a minute may need up to 15 or more pickers that each handle 80 pieces a minute and multiple cameras to coordinate packaging.
“Robotics are not as fast as everyone thinks,” he added. “The robot is limited on chasing an infeed chain on a flow wrapper, so it can get quite complicated when other forms of hard automation do it easier.”
Early robotics relied on vacuum technology that’s simple and effective, but they have drawbacks with handling baked goods.
“The downside is robotics often damage the product or, if you were picking an item like a croissant or artisan product with a garnish on top, it became very difficult to pick it up,” Mr. Watson said. “You were also vacuuming a lot of that garnish back into your system, so it needed to be cleaned out on a regular basis, especially when dealing with allergen-type toppings.”
While the new finger-like grippers rely on sensors to pick up delicate products without damaging them, Mr. Watson added that the technology is pricey and often requires significant testing before implementing.
That research is something that Rick Hoskins, chief executive officer of Colborne Foodbotics, highly recommends.
“EOAT is quite possibly the most influential component to automation solutions successes, but the design process for EOATs can be a very lengthy and difficult challenge,” he observed.
Mr. Hoskins said Colborne Foodbotics uses a proof-of-concept methodology when designing EOATs. This means the company employs 3D printing technology to design tools that are monitored on its robotic testing cells.
“In many cases, we take these robotic testing cells to our customers’ facilities and run various tests with actual production quality products,” Mr. Hoskins said. “This is critical to ensuring our EOATs are designed to the key attributes when producing quality products.”
This process allows bakeries to see in real time how particular EOATs affect the appearance and quality of their products.
“There are pros and cons to many typical styles of EOATs: vacuum, needle or gripper,” he said. “While some may be the most efficient method of picking and placing at high speeds, they may not be ideal for maintaining a product’s shape or quality. Through different testing methods, we can prove these concepts out and settle on the best EOAT for the application.”
Aaron Donlon, product manager, Epson Robots, said 3D printing allows tooling to be customized to fit a product’s needs. He added more advanced materials enable EOATs to flex and conform around irregularities in handling packaged food products.
“Designing active tooling is an effective way to build different pack patterns on a production line,” he said. “Active tooling can enable the grippers to change the pattern or vacuum tooling that is zoned off is another way to create different patterns. Both are used in conjunction with the robot motion to design a variety of pack patterns.”
Felix Pang, sales application specialist, ABI Ltd., suggested soft grippers are the most flexible choice for product handling since they can pick up the widest range of items. The silicone fingers mimic the human hand, and the lightweight EOAT works well with the delta robot — a type of parallel robot that consists of three arms connected to universal joints at the base — as a faster option.
“From a programming perspective, we can dictate how much the fingers open or close and the pressure of the hold,” Mr. Pang said. “This capability makes handling product of different sizes easier. Using machine learning, we can teach our robotic systems to handle a variety of products, including loaves, buns, pretzels and cakes.”
Robotic versatility has improved over the past few years. Mr. Hoskins said it sometimes means using one tool for several different sizes and varieties of products.
“However, we never want to sacrifice overall system performance only to allow for more flexibility,” he added. “In these instances, we will use the auto tool change feature of our systems that allows our customers to change over to an EOAT with the push of a button. This brings a tremendous amount of flexibility to the system but optimizes the performance of the system across a wider variety of products.”
This article is an excerpt from the June 2022 issue of Baking & Snack. To read the entire feature on Robotics in Packaging, click here.