Page 134Journal of American Baking www.asbe.org

I would like to start by thanking ASB and Pat in particular for including us. This is my first conference and it has been a great learning experience already so I definitely appreciate the opportunity.

(Slide 2) I am going to jump right in and start with a little audience participation activity. Just by a show of hands, who knows what that tool in the picture is for? That is the answer I expected. Now, who thinks they have an idea of what that tool might do? (Slide 3) Any do-it-yourselfer who works on and has tried to do his own plumbing job knows if you are going to work on faucets you had better have a faucet puller. But you do not know what that tool does until you see it in context and you see what it is used for. That is what this presentation is about. (Slide 4) If you have only heard about robotics, not really sure what they can do, we are going to go through a bunch of application videos of customers in your industry who have successfully applied robotics to their operations to improve productivity, reduce cost, in some cases improve quality. We will show some of those successful examples and hopefully as we go through there would be something that you can relate to that looks like a process you have in your facility. It will make sense and maybe give you some ideas on how to take advantage of the technology.

Everything you are going to see in this presentation is really downstream of everything that Ravi and Jim talked about. We are going to be down at the end, where now you have a discrete something, a discrete product that can be handled, a primary product or the product in the primary packaging, the secondary packaging and so on all the way through end-of-the-line palletizing. Historically that has been where a lot of customers start. (Slide 5) So here is a typical robotic palletizing system. This one is actually relatively high speed. If there is a perception you have that conventional palletizers can outrun a robot, it depends on a system configuration. (Slide 6) Here is a palletizing system where the cases are marshaled into a picking station, they are collated together in a group and then the robot picks and places the cases using vacuum. What you will see, that as the

Robotics in the Bakery

Dick Motley is going to talk to us a little bit about robotics in the bakery.. He is currently Account Manager for FANUC Robotics’ National Distribution Sales group. In this role Dick provides sales, marketing and technical support to FANUC Robotics’ system integrator partners who focus on non-automotive industries, including food and consumer products. Prior to his sales role, Dick served for 18 years in a variety of engineering capacities within FANUC Robotics, including leading the specification, development and introduction of several of FANUC’s robot, software, and controls products. Before joining FANUC Robotics, Dick spent 5 years with General Electric in an advanced manufacturing automation systems group. Dick holds a B.S.M.E. With Distinction from Pennsylvania State University and jointly holds several patents for innovations in automation and robotics technology.

By Dick Motley

Page 135www.asbe.org Journal of American Baking

robot builds the stacking pattern, there is an actuation also built into the tool to create that spread that you see that is required by this particular stacking pattern. When the robot picks this group of boxes, it really has the ability to create through a servo actuation, very different product spacing depending on the product size and the desired packing stacking pattern.

Another thing about robotics and I guess automation in general, but particularly as compared to a manual operation is the strength of the case to protect your product inside is in the edges and the corners. You want edges stacked on edges, you want corners stacked on corners and that is one of those areas where a robotic system that you are doing manually today can definitely help with your quality and reduced product damage in shipping.

You will also notice the robot is running two different products simultaneously. You will see in a little bit how that can be stretched to even including and more products in the system.

(Slide 7) That robot is using a vacuum tooling system. You might say...well, Dick my product does not lend itself to vacuum or there might be some unique characteristics where you do not think that that is going to be applicable. This happens to be a hooded tray and so instead of vacuuming it, where we would just pull the hood off, this particular application uses a clamping style gripper. You will notice that the robot is able to pick from one of two conveyors where the products are delivered either lengthwise or widthwise. The robot can selectively grab whatever number of cases it needs again to create the desired stacking pattern. Then you see all the other same attributes of corners on corners, edges on edges and really protecting the product integrity during shipping.

What you have seen so far is just a couple of SKUs running in a system. You might say...hey, Dick, I have a lot of SKUs I want to run in lower volume. Here is an exercise in making the absolute most of the available reach of the robot. Here you have a single infeed with a variety of SKUs, eight different ones, flowing into the cell and the bar code is read at the entry point into the cell and the robots are sorting selectively to one of the eight different pallets. In addition the robot handles the bottom slip-sheet that is put under each load. So really bringing a lot of functionality out of this one

Robotics in the Bakerymachine and that is a theme you will see consistently as we go through some of the rest of the videos is some of the best robot applications are the ones where the end customer or the integrator had the creativity to imagine a lot of things that they could do with the robot. That is where robotics will really stretch its legs versus conventional automation.

(Slide 8) I have heard a lot about ingredients while I have been here and I know a lot of ingredients are shipped in bulk, but I have also heard that a lot of ingredients are shipped in pallet loads similar to what you see here. This happens to not be a bakery product, but the principle still applies. You can see that we have a very different looking tool now out on the end of the robot. It is kind of a claw thing that supports and then captures the bag and builds a very tidy load that ships very well. In this particular system you see that it is a relatively high-speed system capable of over 40 bags a minute. One of the ways that is achieved is the elevation, one robot is a little higher than the other, so they work collaboratively where one robot finishes kind of the lower layers of the load and the load is indexed, the other robot finishes the top layers. That is another strategy for getting high throughput if it is needed.

(Slide 9) Nasty place for a person is a freezer. I am imagining some of the kinds of products that Jim just talked about coming out of the sheeting system that then get frozen and you want to keep them that way and maybe through the palletizing process. You get in a deep freeze, -20 or so, and all of a sudden you have a real ergonomics issue. Typically it will take three people to staff a single operation just because you have to spell them to kind of give them some relief from the extreme cold, so that is a great job for a robot. In our industry we talk a little bit about the 3D’s: anything that is dirty, dangerous or difficult is a good job for a robot. This probably has at least two of the D’s, it might not be dirty, but it is definitely difficult and cold definitely presents some challenges in terms of protecting the personnel.

(Slide 10) Obviously these are flour bags. These are retail packages at this point. This particular product gets shifted into what is kind of a notoriously bad ergonomics problem in a lot of industries and that is a high-walled container that requires whoever is loading it, almost by definition, they have to do a lot of bending, reaching, stooping that is really tough from a

Page 136Journal of American Baking www.asbe.org

repetitive strain perspective. So here, not unlike what you saw earlier with the cases, the bags of flour come in, they kind of get marshaled into position, there is that reciprocating crowder that puts them in an array, robot grabs then and then stacks them in the gaylord, again using vacuum in this case.

(Slide 11) Here is a lot larger robot handling lids for baking. Cannot say I have seen it, but I am told there are conventional types of equipment that will do this similar function. Except this particular customer had two types of lids that they needed to run. Being able to buffer between the processes, switch out between the lids, all of a sudden the cost of that conventional solution and the floor space impact started to creep up and up and up to where a robotic solution was a better alternative. You can see that the robots selectively, depending on what is coming in from the back end of the ovens, then before it goes to where the dough is put into the pans, depending on the demands and the ebb and flow it is buffering between those two processes. In some cases doing an intelligent thing where you go back and forth from the storage location or if there is some unexpected demand downstream, it will immediately transfer pans to go right back into the process. It can load to and from dedicated racks or it can use the in-process carts that are typically used for the lid transportation.

Here is one of those videos that I think really is almost a poster child for that creativity that I was talking about a few minutes ago. (Slide 12) Here is someone that is packaging a variety of different cookies and crackers. If you watch the video and notice how many different product types go through this system, there are some carton products, sleeve products, and some products that are in tubes. All of those are accommodated by the same system configuration with virtually zero changeover. That center robot is kind of a traffic cop handling empty cases from the case erector and then pulling full cases out from the packaging stations, also for the open top tray product that you see here. You are going to see it goes to grab the lid, there are a couple of adhesive nozzles here, and you can see that the robot peels the lid, passes it over the adhesive, places the lid and then that is glued in place. There is a lot of functionality out of that center robot. It is doing a lot of different functions in a compact space, again with really easy changeover among all those different product styles.

(Slide 13) Here is an application using a robot to pack buns and rolls. You will see another video very similar to this and a few slides also handling bread. It was really a compelling case for a robot in this sort of application due to the different packing patterns that you can create with a flexible robot. All different product configurations: 2x4, 4x, whatever the different configurations are. The robot can create the desired pattern on the fly and can load either the trays or the baskets or, you will see in a minute, can switch over and also pack corrugated shippers. There is also one other function that you will see in just a second here. After it creates the layer of rolls and then places them into the box, you will see that it also grabs the U-board that separates the layers. You are looking at a bunch of different interface points, you have in-feed conveyors, an output location for baskets, an output location for boxes, and you have the U-board storage. That would take multiple devices if you were going to go with a conventional approach. Really in this case for this customer, made it possible to justify financially pursuing this automation. (Slide 14) This one is a grain-based product. One of the packaging options is a single serving bowl. The bowls have a taper to them, so from a shipping efficiency perspective it is advantageous to alternate the taper to nest the bowls. You can see in this case, the robot grabs half of the pattern. There is a supplementary device that picks every other bowl and inverts it and then the robot grabs the second half of the pattern in the inverted configuration and then case packs four cases at once to achieve the throughput which in this case was over 300 bowls a minute. (Slide 15) Here is a grain-based bar product. One of the things you will see here is that the robot is working with what I will call a racetrack or there are a couple of different styles of collating these kinds of products at high speeds. You will see kind of like that palletizing gripper you saw a little while ago, this has an expanding and collapsing function to go from the spacing in the collator to the spacing that goes into the carton. The robot also acts as the transfer device from the cartoner out-feed and then to the closing in-feed downstream. In the case where the product speed was too fast for a robot to handle, there really was not a reasonable automation solution for the product. (Slide 16) This customer said...okay, I have this high-speed operation; I will have the robot handle my packaging materials.

Page 137www.asbe.org Journal of American Baking

They had a cartoner that was mission critical and they said we do not ever want that magazine to run out, we do not want to have to staff that part of the process. We burned through enough of these cartons that we want to invest in automating that because we recognize, that overall, we will get higher performance out of the system if we do not have interruptions due to the cartoner running out of material. (Slide 17) Here is a muffin depanning application. The robot is extracting the muffins with, it almost sounds like an oxymoron, but a gentle piercing function to grab the muffins out of the pans. With all the different toppings that can go on these products, there really was not any other viable handling solution, so this works like a champ. The kind of needles that go in and grasp the muffin, go in a particular direction to almost cradle it during the transfer. The robot transfers add up to 22 pans a minute into the retail clamshell package and then through an automatic tool change can also pack into a bulk tray format. That is another attribute of this system. Now we have gone from handling packaged product to handling raw product. The sanitation of the end-of-arm tooling on the robot becomes a key item and so the tool change also facilitates doing the clean up. (Slide 18) This probably looks a little bit familiar. This is very similar to the system that does the buns and rolls except now it is doing bread. With bread there is really a compelling case from a shipping efficiency perspective. The shrinkage that occurs from when the bread is warm to after it ships and gets on the shelf. There is a tremendous efficiency to be gained in terms of making the most out of the space available in that basket. Applying just the right amount of compression to fit the maximum, not the maximum, but the best amount of bread into the basket helps both with all the logistics associated with the basket handling. If you think about it, if you have more baskets in process, more people handling them maybe elsewhere in the system, more trucks that are trucking the baskets around, there is a huge gain maybe as much as 30% from a logistics efficiency perspective. Also, here is where you get at the quality aspect of what robotics can do. Now you have less product damage because it is kind of snugly fit and it does not rattle around as it shrinks in shipment.

I just learned from Jim’s presentation, I think these products come out of a sheeter. (Slide 19, 20) What you will see in this system is these are a variety of frozen dough products. This first one is kind of an individual serving size pizza. You will notice that the robot is grabbing a complete layer at a time and then also handles a separator sheet that is introduced into the case.

What we have introduced now is vision as the eyes of the robot. As all those products flow down a conveyor, the robot is actually able to find where they are and then the control mimics hand/eye coordination and lets the robot kind of pick the products gently off a moving conveyor. You can see as you watch this video, tremendous product flexibility. All those neat, different shapes that Jim just showed us his sheeters can make. We just went from the small single serving round to a much larger multi-serving round. Now in that same system, with just limited changeover to the grippers on the end of the robot, we are doing a breadstick configuration. Introducing vision which is really rocketing in terms of acceptance because of the way that the performance has increased, the user friendliness has increased and the cost has come down, it is kind of hit that tipping point where we see a lot of adoption of vision in the industry for applications like this. Similarly here is product in bags or pouches. It seems like my wife is always bringing home the 100-calorie packs or small servings of small muffins or cookies. As you watch this system run, again this is enabled by vision. You can see the kind of signature red LED lighting upstream. There is a camera in that area identifying where all the products are and there is some pretty sophisticated software that is acting as the foreman for these three robots. Instructing, you get this one, you get this one, and you get this one. Then another benefit of vision you will see in just a second here. There are a couple of bags that are coming through that didn’t get slit for some reason in that one bin and the vision system just screens that out and ignores it. There is an advantage versus maybe a conventional technology where that might cause a jam, cause you downtime. You have to take the line down and go and correct that and here you have a system that provides all the changeover flexibility you want and also kind of protects itself from nuisance downtime.

Robotics in the Bakery

Page 138Journal of American Baking www.asbe.org

(Slide 21) Here is a waffle product. Now we are squarely into the domain where vision is really a key enabler for being able to do these kinds of projects. This is a pretty high rate, over 600 waffles a minute coming at us. You will see in this video as well, not only is the vision system directing the robot to these waffles, but also you will see that it is similar to the situation with the bags that did not get separated. You will see that there is some product on the conveyor that has pretty serious defects that you do not want to ship to a customer, like that one. The robot is basically screening for defective product, do not send that to your customers, and also eliminating that as a potential cause of downtime because of the irregularity of the product. Here is that supervisor software, that foreman software that I told you about, that is under the hood here making that sort of application possible. (Slide 22) Here is maybe one of the most extensive uses of the technology. This is a frozen specialty product that comes out of the freezer and then has to get transferred to the in-feed of flow wrappers. A key issue here is dealing with the exception: it is easy to define what the system is supposed to do under normal operation, but now how do we deal with the fact that the supply of film to the flow wrapper is not infinite and sooner or later we have to resupply that? You have two product infeed conveyors going to each of three in-feeds to flow wrappers and you have to plan that you are going to have to resupply the film. On a normal operation, of the 18 robots that are in the system, six apiece feed the three out-feeds. 18 divided by 3, right? But now if you have to take one of those flow wrappers down, this system automatically switches over. The speed of the remaining two flow wrapper in-feeds is increased and the robots rebalance and instead of six per 3 lanes, they go nine per two lanes, all automatically. You have tremendous production flexibility, tremendous efficiency of that line. I read a lot in trade publications about how important OEE is becoming from a plant performance and even a sustainability aspect that we have the plant turned on, we want to be making product. That is really an area where robots can shine. They can reduce downtime as you have seen with some of the vision applications, also changeover flexibility where you eliminate lots of change parts, think of the frozen dough application. Here you are seeing where the robots automatically compensate for something that is given is going to happen in production. That one is last for a reason. That

is a pretty graphic illustration of some of the principles that we are talking about. (Slide 23) Pat wanted me to comment on what is next in robotics. Simulation technology is not necessarily new, but the features that we keep packing into simulation software to make it very intuitive, very user friendly for integrators to develop system proposals, test drive the virtual factory, the robot system, really tax it from, again, not just what is supposed to happen under normal operation, but how does the system respond to exceptions or special conditions. If you do things like vary the randomness and spacing, you can create surges and blockages and see how the system responds and just watch on the screen how that system will behave when it is subjected to those kinds of variations in production. Simulation is just growing by leaps and bounds and really gives you the confidence when certainly there are other details to be worked out in terms of getting a functional gripper that handles the product reliably and so forth. In terms of overall system performance, you can really gain a lot of insight how the systems behave under both normal and exception conditions. Another huge area of advance is in the area of safety software. The robots now are almost self-aware. They know where these imaginary volumes around the arm are -- actually representations of what the safety software is calculating in real time. The whole time the robot is moving, it is aware of where those volumes are and those can be tested against virtual boundaries. If there are two problems in a robot system, one of which is “keeping the lion in the cage” and the other is “keeping people from going in the cage with the lion”, this one helps with that former challenge of keeping the robot in a compact cage that fits your floor space efficiently and is very intuitive to the operators that they understand. Those are the limits of where that robot can go because you typically will put up a guarding system, a perimeter guarding system that makes that virtual limitation very real and tangible for the operator. (Slide 24) Another thing that is pretty exciting is we’re moving from the manufacturing domain, which is a lot of what you have seen so far, and moving into warehousing domain. Building store-specific, truck route specific, even aisle-specific loads. You think about that challenge, that is really creating a three-dimensional

Page 139www.asbe.org Journal of American Baking

puzzle, right? If you think about a person doing that, you would say this fits better over here, this fits better over there. There is a lot of brainpower going on to address that challenge. There are some really exciting software and hardware tools that allow us to handle a wide variety of products and then emulate those mental gymnastics of creating the three-dimensional puzzle in a warehouse setting. I commented on vision already. Vision is just exploding both as a guidance technology to give you a sighted robot, again to emulate hand-eye coordination, but also to do in-process product inspection as you saw on some of these. In some of the early applications when customers were looking at labor substitution with robotics, it always got to that point where it dawned on the customer, that guy is not just moving the product around. He is also giving me some in-process inspection that I do not want to lose and vision can help retain that. Because of the adoption in the food industry, there is also significant growth in the number of food robots, the actual mechanical arm that goes out in the environment and is in proximity to the food and there are a lot of different varieties. Some of them are adaptations of existing industrial robots. Some are ground-up, purpose-built designs with a very sanitary design that resists some of the sanitation chemicals and in some cases, as at least a starting indicator for cleanliness, have achieved USDA acceptance. (Slide 25) So let us say that you saw something among those videos that makes you think...you know what, maybe I could do that in my plant. What would I recommend that you think about if you were to undertake a project? Well, first is identify the right opportunity, and the key here is do not look just at labor savings. There are so many examples that we have of customers who undertook a robotic automation program where they fixated on the labor savings initially. Then they got into the program and they realized that was the tip of the iceberg. The real benefit was schedule consistency, production consistency, some of the things that Ravi talked about in terms of thermal sensitivity to the kinds of products you folks make and how difficult that must be for you folks to maybe deal with shift change or operator going to the bathroom or whatever the case might be. Those are the kind of things that customers are looking at to really bake in (no pun intended).

Really consider all the benefits that might come out of robotic automation from productivity to OEE benefits, reductions in change over time, reductions in change parts, reductions in maintenance. Industrial robots have a really stellar reliability record that was really driven by some of our roots in the automotive industry where downtime was just so incredibly expensive to those types of customers. Another key to success is really making sure that all the stakeholders understand what you are trying to do and why. That means different things to different people. To your senior management they probably want to understand the financial impact. To your maintenance folks, they probably want to understand how the equipment is going to be maintained, and in a lot of cases that will be a positive. Our typical industrial robot might need to be greased twice a year, so very low maintenance burden in terms of the basic motion platform in the robot. What you get back is, we have meantime between failure statistics that are up in the 90,000+ hour range. Just bulletproof for reliability with low maintenance burden for the basic robot platform. The key is to make sure that everybody understands what is in it for them. Because we have seen lots of times where if that message is not clear, and the equipment lands on the plant floor and it is not embraced by the folks that have to make that run production every day, it just will not be successful. We have a technical offering, a pretty high-tech product, but to me this is more key than any technology we offer in terms of an indicator of success. Another key factor for success is finding the right integrator. The videos you saw are by a lot of folks who speak “bakery” first or “food” first and then their second language is robotics. They really understand your process and they studied how to complement their process knowledge with robotic technology to give you the best possible solution. The right integrator is another key factor. If you have a relationship with an integrator, I think the key is to get all the expectations of the system on the table. It can be a formal specification process or just a lot of heart-to-heart talks. Defining what the system is supposed to do under normal conditions is the easy part. The part that tends to take a lot more time and is often subject of a lot more discussion is how will this system deal with exceptions? What happens when a

Robotics in the Bakery

Page 140Journal of American Baking www.asbe.org

bad product comes in the system? What happens when the upstream operation goes down? What happens when the downstream operation goes down? What happens when I have to take one robot off line for maintenance or something to do in the interim? Those are really good discussions to have earlier rather than later because when the system is on the floor and there is not agreement on how those exceptions are going to be treated, that is generally a point of contention. Invest in training. This is closely associated with the “involving your stakeholders”. Part of involving your stakeholders is engaging those folks who are responsible for making that system run on the factory floor day in, day out. They have to have adequate training so that they understand the technology. They can make it successful and training is again another key indicator that folks understand what is coming, they know what to do with it when it gets there, they understand the benefits and they know how to run it successfully. The last one, this might seem trivial, but when you have a successful system, do not sweep it under the rug. With some companies we have seen kind of an infectious enthusiasm that hey, that was not too hard. We had maybe a good relationship with an integrator we trust and let us broadcast that and find additional opportunities for some of those benefits that I have talked about along the way. I do want to thank the different folks in our integrator network that provided the videos that you have seen so far today. You will find that most of the robot manufacturers do not speak bakery. That is not our first language. Our first language is robots and our second language is probably automotive if you trace our histories back. These folks are the guys that speak bakery and food first, their second language is robots. These are great folks to work with if you want to move forward. (Slide 26) So at that point I will field any questions you might have.

PAT WILKENS: Before we do, I wanted to say one thing that I neglected to do. Obviously, Dick is extremely knowledgeable on this subject and presents himself very well and there is a good reason for it. Dick is a National Account Manager for Fanuc Robotics. Dick

graduated from Penn State University with his BSME with a Distinction. He has 25 years’ experience, which obviously shows, in industrial automation and robotics. He also jointly holds several patents for innovations in automation and robotics technology. I did not want to exclude that information which my nerves got the best of me in the beginning and I kind of rushed out. But are there any questions for Dick?

UNIDENTIFIED SPEAKER: The level of operator education...I think one of the things that scare people about robots is who is going to take care of it? Who is working in my bakery right now that is going to support this?

DICK MOTLEY: Good question. The question was and I will just repeat it for the cameras. What skill level do we have to assume for running a robotic system? And honestly, the answer is another discussion I probably should have touched on which is what is your long-term support strategy? As far as your operators and what they need to know, it can be reduced by an experienced integrator. Their interface can just be a bunch of buttons or touch screens just like they do on every other packaging machine. You have sophisticated machinery, baking and packing, already in your facilities, so that part of it does not have to be rocket science. When you drill down, you want to deal with the robots specifically. There you have a couple of options. You can train up to be completely self-sufficient. You can train up to the point where you can take care of nuisance stuff like a fuse blown because something went wrong on a maintenance procedure or whatever. You can also train up to be completely self-sufficient and tear down about every nut and bolt if that is what you want to do. What a lot of customers decide to do is, train up to the point where the guys can do the basics and then they rely either on the integrator or the robot manufacturer to provide the rest if they get into a deep problem. There you just have to go in knowing that if I hit the really rough terrain, I am going to have to have a contingency for that 12-36 hour window where the response team is able to get onsite and take care of the issue.

UNIDENTIFIED SPEAKER: As far as your robot suppliers, are they stacked and have enough people to service your customers?

Page 141www.asbe.org Journal of American Baking

DICK MOTLEY: I cannot speak for other robot manufacturers, but, yes, I would say Fanuc Robotics is. We have a pretty significant commitment to after-sale support. We have a toll free hotline. By the way, the question was: Are the robot manufacturers staffed to support a lot of customers? To give you a little appreciation for our customer base, in the Americas we are supporting an installed base of over 100,000 machines. We are able to do that partially because the reliability is so good, but also because we have staffed up appropriately to provide the service.

UNIDENTIFIED SPEAKER: So how many guys does it take to support 100,000 machines?

DICK MOTLEY: Well, we have, and I am not supposed to get into this commercial discussion but we actually have a significant presence at our headquarters in Michigan, where I am located, and then also a number of regional locations.

UNIDENTIFIED SPEAKER: I was hoping you would say it would take like two people.

DICK MOTLEY: Think of the Verizon commercials where the network is behind you. You can just see the guy in the front saying you are good, or we have that strength behind us if you need it.

PAT WILKENS: Any other questions? Well, Dick, I would like to thank you. On behalf of myself and ASB for a terrific paper.

DICK MOTLEY: Thank you.

Robotics in the Bakery

Page 142Journal of American Baking www.asbe.org

Slide #1 Slide #2

Slide #3 Slide #4

Slide #5 Slide #6

Slide #7 Slide #8

Page 143www.asbe.org Journal of American Baking

Slide #9 Slide #10

Slide #11 Slide #12

Slide #13 Slide #14

Slide #15 Slide #16

Robotics in the Bakery

Page 144Journal of American Baking www.asbe.org

Slide #17 Slide #18

Slide #19 Slide #20

Slide #21 Slide #22

Slide #23 Slide #24

Page 145www.asbe.org Journal of American Baking

Robotics in the Bakery

Slide #25 Slide #26