In the Trenches with Collaborative Robots

P & G’s Mark Lewandowski walks us through a risk assessment for successful robot applications from an engineer’s perspective.

In the Trenches with Collaborative Robots
In the Trenches with Collaborative Robots

According to the latest research from PMMI Business Intelligence, while many CPGs are using robots (mainly for palletizing) most have not yet employed a collaborative robot to work alongside a human.

P & G, maker of everything from Tide to diapers, paper towels to Braun shavers, started investigating collaborative robots back in 2010! And applications as varied as those products listed above all use robotics as a key way to automate the process.

“If you haven't figured out from the way I'm dressed yet, I am not a marketing or a sales guy: I am an engineer. So, what you're going to get here is the perspective from an engineer in the trenches,” Lewandowski told the crowd at the FORUM on the floor of PACK EXPO Las Vegas.

“So, at P & G we do pioneer. We believe that through our engineering organizations, and through our process and R&D organizations, we bring a lot of innovation into how we manufacture and create products,” said Lewandowski.

“So why did we, as a company, start using collaborative robots? Well, they’re quite simple to implement and use. The cost of implementing some of these cage-free robots can be a part of it…but, a bigger reason is the reduced engineering costs: being able to easily integrate and develop an application using collaborative robot technologies and greatly reduce the amount of time and effort needed, which reduces the amount, the overall cost, of a deployed solution onto your factory floor,” Lewandowski continued.

“We can take these robots that are adaptive, they’re reconfigurable, we can take them from a certain application today, very easily reconfigure that application, and have it doing something else tomorrow.”

So, collaborative robots are smaller, lighter, easy to move around and have intuitive programming. But Lewandowski disputes the “sales claims” that the robots are “safe” and require no fencing or guards.

“We hear that all the time, that these robots are cage-free and require no fencing or guards. Well, not necessarily true,” said Lewandowski.

Lewandowski points to ISO/TS 1566 as your guide. He was on the committee who wrote the safety standards for collaborative robots. But before these standards were written, P & G approached collaborative robot applications from the concept of safety first.

“How do we make sure that whatever we do with collaborative robots are going to be safe? Because, as it is with any new technology, you don't want to do anything early on that's going to give it a bad name or reputation; and if we had any incidents or anything that went on with a collaborative robot as we were trying to deploy this new technology into P&G, specifically, and into the robotics industry in general, then that was going to be a very difficult barrier to overcome if we had any incident,” Lewandowski told the Forum crowd.

He continued, “So, we've got to go in with a mindset of safety first, and I think as you look to deploy applications into your operations and into your sites, go in with that mindset. Make sure it's safety first, because this is a new technology. People are somewhat afraid of this technology, maybe, in some cases, and so you need to go in there with a concept; here's what we're doing to make these applications safe. And being able to show them what you're doing and how these applications are safe. So, go in with that concept of safety first.”

The best way to achieve this, said Lewandowski, is with a risk assessment. “Now, the first thing is follow the standards. There are guidelines and directives in the standards. In the US it's through the RA15066, and the international standard is the ISO10218, which are the robot safety standards. They're identical documents.

“It's risk assessment. It's risk analysis. Making sure that you understand what the hazards are in that application, and then what are the safeguards and things you could put in place to deal with each of those hazards that you identify? So, risk assessment is a key requirement, not only in all the robot safety standards, but in just about any safety standard that's been published out there for any type of machine,” said Lewandowski. “It's the concept of risk assessment. That is the key requirement to make sure that you develop a safe application. You cannot have robots and humans working together if you don't do a complete, comprehensive risk assessment. And it's not just the robot, and that's one of the things I think is one of the biggest issues that we deal with when people are first looking at collaborative robots, is they hear ‘we have this safe, collaborative robot.’

“But we're not deploying robots. We're deploying applications. We need a robot that does something, okay? Take an arm and make it move back and forth all you want, but I can't do anything with that. What do I need to add to it? I need to add an inner-arm tool; I need to add some ancillary equipment that goes around that to bring products in or take products out, or whatever I need to make that application work.”

That’s why Lewandowski emphasized that the overall cell must be considered. He put it this way: “It's not just a robot, it's an application. So, when we look at developing a safe robot it really is developing a safe robot application. So, we need to make sure that we include all the things: the tooling, the grippers, and all the peripheral equipment around it that's part of that application. And, in many cases that's where we run into bigger issues, not with the robot itself.”

Lewandowski pointed out that P & G had a great need for what’s called a “co-pal,” or a collaborative robot palletizer…a simple collaborative robot picking boxes off the end of a conveyor. “It's probably the simplest application we could figure out what to do with a collaborative robot…picking boxes off of a conveyor system and stacking them into pallets. So, this was done in our corporate engineering labs, our robotics lab. I think this was 2013, maybe, when we first did this, so this was almost six years ago.

“And so, we were really happy. We're the engineers. We get out there, we take a look at this, and we say, ‘Hey, we've got this application. It works great. It can take lots of different cases, lots of different sizes, and we've got this application. We're ready to roll this thing out.’ But then we got our robotic safety people involved, and they started taking a look at it and said, ‘Well, have you done a risk assessment?’ We're like, ‘Yeah, we've kind of looked at it, but we haven't really done a full risk assessment.’

“So, if you were going to do a risk assessment on this, what are some of the hazards you might see as you see kind of what this process and application looks like? The robot arm itself? It has force limiting. But what the robot doesn't necessarily take into account is maybe some of the other hazards associated with other parts of the application.

“The end-of-arm tool, for example, or the box that's being moved, or what about the interaction of the robot and going to pick up the box off the conveyor? Are there any hazards associated with that? When you start looking at collaborative applications, there are new hazards that might be present in a collaborative application that aren't present in a traditional robotic application.

“And why is that? Because in a collaborative robot application there is a potential, if not the expectation, that robot and humans may come in contact with each other, and that's really the difference in a collaborative application.”

In a traditional application, said Lewandowski, great care was always taken to make sure that robot and human were kept safely apart. But in a collaborative application you need to be concerned with things like can the robot have a negative impact on the person? The gripper as well as the part that's being carried by the robot all now potentially can come into contact with a person, so you've got lots of different potential impacts, as well as crushing potentials or pinching, not only potentially in the robot itself, but between parts and fixtures, between the arm and fixtures or between the arm and other things that might be present in the cell.

“So, if I just looked at the robot by itself I may not consider those type of hazards,” said Lewandowski. “But as I look at the overall application and everything that's part of it, these are all hazards that I need to look at and consider in my risk assessment, and make sure that I've adequately addressed in order to make sure that the application is a safe application. And then you've also got more hazards around the arm itself, and things around the end-of-arm tooling and the gripper, and in many collaborative applications the gripper is often the part that people overlook or don't consider, or certainly don't look at close enough. Those are certainly things you need to look at.

“Padding can be a great way to minimize forces and pressures that you may end up with in potential contacts or collisions between a robot and a person. As well as collision detection. So, not only just padding to soften the blow, but can you actually detect that collision, not using the robot arm but using some type of external collision detection to actually stop the arm, which can greatly reduce the force? You can use geometry and limits, using space to your advantage. Design things in such a way that the crushing and clamping hazards aren't possible because there's just not enough space, or not enough minimum space for those to occur.

“Use the envelope limiting, the safe limiting features of the robots, to limit where the robot can go, so that it can't get into a position where it might create some of these hazards. You can also use conditional, or what's called geometric limiting of the force that will proceed. There are safety features, depending on the brands of collaborative robots you use, that can limit force in a certain direction or speed in a certain direction. Take advantage of all those built-in functions as a way to limit the movement of the robot so you don't create any hazardous situations.”

P &G is making at least selective use of more traditional means of guarding, even collaborative robots. That means using traditional guarding, or using laser scanners or light curtains to limit access to areas that may be considered a higher hazard area but still give you access to the application to do that.

“One of the things around power and force limiting is the concern about the head and neck region of the body” said Lewandowski. “That is a very high-concern area. If you ever looked at 15066, it's an area which they don't provide a lot of guidance because they tell you there's just no way to make that safe.

“So, you really must design your application in such a way that you are limiting or completely minimizing as much as possible potential contacts in the head and body, the head and neck region of the body, because there's just not really a way to do that where there's a safe contact. So, you need to be very cautious about that in your applications, or develop other ways to deal with that, which at P & G includes active padding, so rather than just using it to spread out the force this actually detects the collision and stops the robot.”

What it really comes down to is that from the time you develop an application to before you roll it onto the floor, you need to go through this process of risk assessment. “You need to go through the process of validation,” said Lewandowski. “You need to go validate your safety systems and your safety functions to make sure that you're meeting all the requirements, that the forces that occur in there are below the allowable limits, and that requires a lot of work in many applications.”

 

Make plans to visit PACK EXPO East in Philadelphia, March 3-5, to see on-trend packaging machinery and materials.

More in Robotics/Vision Systems