Automation in the Age of COVID-19 – Advanced Manufacturing

Collaborative robots are easier to implement in part because they can be trained by manually dragging the robot head where you want it to go.

The longstanding trend toward manufacturing automation has understandably been accelerated by the COVID-19 pandemic. Roughly a year into the crisis, its a good time to ask about longer term impacts. Has the pandemic pushed automation into new areas? Have suppliers made automation more flexible? Easier to implement?

By their nature, collaborative robots (cobots) are easier to implement than their traditional cousins. Because cobots are designed to share a workspace with their human counterparts, as a matter of routine, they often dont require guarding and can fit into relatively small, occupied spaces (when confirmed by safety risk assessment). Theyre also easier to set up, said Dick Motley, the Director of FANUC America Corp.s authorized system integrators group in Rochester Hills, Mich. He explained that a user can, in part, train FANUCs CRX series cobots by literally grabbing the arm and leading it around. This makes setup of a simple application really straightforward and intuitive. Youve got robotic automation that can be deployed quickly. He added that theres a growing ecosystem of peripheral suppliers for grippers and pedestals for the robot to sit on, and different provisions to easily address the utilities that go out to the robots end-of-arm tool.

The bad news is that cobots are slow. Even though theyre built with sensors that limit the force theyll impart if they come into contact with something, its pretty tough to meet safety standards at high speed, explained Motley. Because regardless of how sensitive your contact sensing system is, youre trying to defy physics if youve got something moving really fast and then need to immediately bring it to rest. So although you might think cobots would be taking over the world of automation (and COVID-19 era sales have been explosive), their applicability has limits.

Motley referred to a relatively low-speed palletizing operation at the end of a customers manufacturing lines as a good fit for a cobot. They were making two to four cases of product a minute, with an incredible density of conveyor lines feeding these products down to the end to be palletized. They didnt have the physical space to do a traditional robot implementation with the stopping distance calculations and all the things that go into a traditional robot cell. Their only automation option was to put in a cobot. If you go slow enough, a cardboard box is probably not going to hurt you.

This is not to say that traditional robots cant operate in proximity to people, or that theyre very difficult to set up. Or that slow-moving cobots can operate without guarding if theyre handling something sharp or otherwise dangerous. To address these concerns, FANUC and other OEMs have systems that restrict either the motion range or the speed of the robot to allow for intermittent interaction with a human, explained Motley. FANUC calls its safety architecture Dual Check Safety, or DCS. Maybe you want to establish a keep out zone on one side of the robot while an operator loads parts or something that the robot is going to retrieve. You enable a software constraint to keep the robot from going there, typically reinforced by a light curtain or a safety mat or a scanner. But then, once the person leaves that loading zone, the robot can be right back up to full speed.

Motley also pointed out that adjustments to DCS can be done in FANUCs Roboguide offline programming environment simulation package. They can be done with a laptop attached to the robot, or with one of our user interface devices, whether its our iPendant or our new Teach Pendant tablet. The new tablet-based interface is particularly easy to use, said Motley. Its a whole new programming style thats not even close to computer language. Its a drag-and-drop icon timeline.

At Promess Inc., Brighton, Mich., Director of Application John Lytle reported that the pandemic has accelerated his companys effort to make its Electric Press Workstations perform additional functions. Promess had already made it easy to add its units to a production line by making them compact and self contained, with an integrated light curtain to prevent injury to the operator. They automate assembly with sensing that determines if it made a good part, and you can put them wherever you want and adapt to things as they change in the plant. Promess has added ancillary functions and enhanced the information the units are able to communicate to the rest of the factory. For example, in addition to pressing two parts together and confirming that the measured force and travel were as expected and the assembly is therefore good, the Work Station might also take dimensional measurements and pass that information along. This eliminates the need for a separate gaging station.

Performing such measurements requires cameras and/or lasers, and Promess integrates the technology such that the end-user still gets a stand-alone, plug and play unit. As Lytle explained, Were focused on making it simple for the end user, so its not like a big science project requiring a camera technician, a PLC expert, a high-level integrator, and so on. We have a software team of 20 people here, working every day to make it easy. So when the customer gets a Work Station, its already set up. Theyre just entering parameters. He added that cameras can do more than measure parts. They can also be used for part orientation. This enables more complex arrangements, like being able to automatically pick from multiple parts on a pallet. Plus Promess has integrated cobots for automatic part load/unload. The end result, as Lytle sees it, is a multi-function Work Station that simplifies the transfer line and contributes to social distancing, while also sending data to the other equipment in the plant via the Internet or Intranet to make a decision about what to do.

According to Joe Chudy, general manager of ABB Robotics USA, Auburn Hills, Mich., all industries, in large and small installations, are looking for ways to remove people from their processes. The biggest increase in demanda boost that can be tied directly to the pandemicis in medical manufacturing, packaging/logistics, and food processing. The latter two are particularly challenging, given their need for extreme speed in the face of inconsistent inputs. As Chudy put it, its no secret that Amazon cant hire enough people and cant automate quickly enough. The same is true of WalMart and everyone else in that space. But the quantity and diversity of the items you have to pick up and sort quickly forces you to implement some form of AI (artificial intelligence).

Chudy said the food processing industry is also hard pressed and driving automation innovation. He described meat cutting and packing as an inherently miserable environment for humans to work in, with the pandemic only adding to the woes. And the protein (as the industry refers to their product) varies from piece to piece. We asked ourselves if we could debone a chicken. What could we do with the wings? Things like that, recounted Chudy.

Given the inconsistencies of the forms of the protein, meeting this challenge required both a smart camera system and AI to orient the robot grippers. Theres also quick payback on limiting the protein lost in making the cut, added Chudy. So the vision technology you use, with water knives or other techniques to cut this material as close as you can, is a big deal. Learning how the protein is presented to the robot, where the vision system should go, and how you should orient it, all factors into the application.

Long term, Chudy thinks, advances in AI driven by these challenges will also be applied to the metalcutting world. For example, random bin picking capability is improving due to improvements in AI, he observed, as well as de-palletizing and some of the logistic [tasks]. Acquisition times are really what matters in random bin picking. How quickly can I locate that part? How fast can I go get it? Its the same in the logistics market. Chudy believes speed is important and processing the volume of image data required to pick the parts has until now limited those applications. [Now] were seeing those applications flourishing in this market as the technology grows and becomes stronger and less maintenance-intense, Chudy said.

He said fixtureless welding and smart welding are also being studied, in which youre presenting the piece and part to a camera and its deciding how to weld it, and the tolerances, and measuring the gaps, and doing all the things a traditional programmer would do. It works well in the lab, but we have not implemented it in production yet.

Vision technology is central to these systems. And as Chudy put it, that used to make a lot of guys in the metal industry nervous, because vision systems previously required specialized technicians to install and maintain. Now these newer applications are able to self monitor.

Motley also spoke of vision technology as a tremendous enabler that has gotten easier to implement. When you give a robot eyes to be able to adapt to the environment, that enables all kinds of things in terms of reduced cost for setting an application up, reduced changeover time, fewer changed parts, better flexibility, higher reliability, and in-process inspection. He said FANUCs new 3D vision technology, called 3DV, offers a versatile, compact sensor that can easily be built into end-of-arm tooling on a robot. The robot literally carries its eyes around with it rather than having just a stationary sensor, although stationary sensors still have their place. And a 3D point cloud provides the robot with much more information than a flat 2D image, he added. With more information about whats in front of it, the robots control is better able to decide what to do next.

Handling packaging and protein is a lot different than a typical metal working application. So, as Dave Suica, president of Fastems LLC, West Chester, Ohio, explained, gripper technology is changing. We have started going to servo-controlled grippers. A lot of the parts are deformable. A regular power gripper can apply too much force, more than whats needed to overcome the friction factor to lift the part. With servo-control, you go to a position, and then it has an override for how much pressure it applies. More generally, Suica said people are tending toward higher end automation that doesnt require manually changing grippers in order to switch jobs. With smart automation and automatic gripper changes, and computer control versus a PLC, you can make it dynamic. You can switch from part A to part B to part C without any people there. We have systems that run for 72 hours autonomously.

While Fastems is best known for large flexible manufacturing systems (FMS), its handles the entire range of robot and pallet handling configurations. Suica said the pandemic has caused some companies to buy and automate single machines just to quickly re-shore a certain product. But whether Fastems delivers a large FMS or a robot for a single machine, it still has a full line of manufacturing management software, said Suica. It still has scheduling. It still has the capacities to run different parts at different times. Fastems prides itself on being able to integrate with a companys ERP system for predictive and dynamic scheduling. So, as the ERP system changes the requirements, we automatically change the sequence of what part gets made when, such that you maintain your flow without building inventory.

Returning to the packaging and assembly challenge, Motley said FANUCs automotive roots have served it well. Our ability to dispense a bead of sealant to seal up a car body prior to painting, and coordinating that very tightly with robot motions such that when you go fast, you dont get a thin spot in the sealant bead and when you go slow around a corner you dont get a big, thick pile up contributed directly to being able to dispense adhesive in assembly operations.

The increase in automation and social distancing driven by the pandemic has in turn highlighted the need for remote monitoring capabilities. Remote monitoring isnt new, and virtually all control, automation, and machine manufacturers offer such solutions. But Fagor AutomationUSA, Elk Grove Village, Illinois, recently went a step further by accelerating the release of an HTML5 based control architecture. As General Manager for North America Harsh Bibra explained, HTML5 is not only browser based, its consistent across multiple browsers. One person might be using an iPhone. Another might have a Google device. A third person could be using Windows 10 on a laptop. With an HTML5-based interface on the machine, they would all see the same thing in a similar way. HTML5 makes your machine platform independent.

HTML5 provides better mobile access to business intelligence too, said Bibra, along with geolocation. With geolocation, he pointed out, you can limit remote connections to devices that are in specific locations, thereby improving security. For example, you could limit a remote connection to an employees house, but not elsewhere, to prevent access if the employee lost his phone. Whats more, added Bibra, HTML5 is without limits. Depending on the power of the logic you write, or the power of the human- machine interface you create, it can provide the Nth degree of freedom in the sense that the person on the other end can have access to anything. That means you could empower a remote connection to not just monitor activity, but also enter machine commands, like cycle start or cycle stop. In other words, remote control is the ultimate in social distancing.

Read more:

Automation in the Age of COVID-19 - Advanced Manufacturing

Related Post

Comments are closed.