Nate Hartman, Dauch Family professor of advanced manufacturing and head of the Computer Graphics Technology Dept., Purdue University, joins Peggy to tackle the world of smart manufacturing vs. smart factories and what are the actual nuisances between digital, smart, and intelligent. The two also talk about the technologies in a smart factory and how companies can prepare their workforce.
Peggy Smedley: We’ve learned a lot about manufacturing with everything that’s happened with COVID, but I thought it would be great if you could help us understand what smart manufacturing is, versus what is smart factory, because I think it’s a great, interesting way to start a conversation today.
Nate Hartman: Sure, Peggy. One of the things that is important to understand when we talk about manufacturing in this next generation or Industry 4.0 paradigm is, there are nuances between digital, and smart, and intelligent. And even though some people use them synonymously, we’re beginning to see a differentiation between those terms. So, a smart factory, generally speaking, is a facility that’s been equipped to gather information about the products that are moving through it, about the processes that are used within it, and to gather information about the other inputs and outputs that are coming from the facility and leaving the facility. Many people gather this information, because their facility is equipped with sensors in many cases on the machines or on the material handling equipment, maybe even as wearable devices on the workers, all in an attempt to optimize how production is done and how quality is enforced, or store safety is enforced.
And so, when you think about the factory, in some instances, it’s not a whole lot different than a traditional factory, in terms of the building or the equipment that’s in it. However, how people use the equipment, the data they can gather, the analysis and intelligence that can be gleaned from it is where we start to move into smart manufacturing, the manufacturing process itself. So, if you think about historically, we’ve had things like statistical process control, or we’ve had machine monitoring. In many cases, those were alone types of activities. We might have watched our material handling systems. We might have deployed certain types of automation in those environments, but again, in many cases they were disconnected and discreet activities.
And so, one of the hallmarks of having a smart factory and by extension being able to do smart manufacturing is this idea of connectedness. And so, smart manufacturing relies on digital information, it relies on digital networks, it relies on digital sensors, it relies on a number of individual technologies. But when you combine them together, you end up with a smart environment, and ultimately a smart process. And so, just as in the connectivity within smart manufacturing is an important, once you have the information, having a smart manufacturing enterprise also implies that you are going to analyze that information and make decisions from it.
Smedley: So, let’s take a step back. I always like to tell everyone I want to peel back the onion. I’m really big on interpreting all the information we talk about, because that’s where we’re starting to see all these data scientists come in. So, when we say we want to analyze it, we want to have predictive maintenance. We don’t want to be reactive. We want to be proactive. We used to say, when we first started the IoT (Internet of Things) and before that, we had machine-to-machine, okay?
Are we talking about this connectedness that you just described? Are we connecting in the smart world that you’re talking about of smart manufacturing, that factories that are separate, that are not islands of automation will now be connected separately, different buildings? And that’s going to create this smart manufacturing that you’re now referencing, that we’re going to be able to make decisions and be able to do things better, faster, quicker, because of a collaboration that we’re going to be able to work with our partners better, or however that might be to make machines better or to be smart manufacturing machines?
Hartman: So, Peggy, I think that’s a really good question. And certainly, there is a there’s nuance to what it is that you just described. And I think following up maybe on the last point that you made, is this only within our own factory, or might we be connecting buildings, whether it’s across one organization or across multiple organizations? I think the answer is yes, on both accounts. It will obviously depend to some degree on the company and how they operate. Some companies operate, they only have one facility, everything’s wholly contained within the four walls. Other companies are bigger, they may have multiple divisions scattered across different states or different countries. So, the connectedness scales, it scales internally… Or excuse me, it’s applied internally, as well as externally.
And so, one of the things that is important to think about is in terms of this connectedness, many cases, we’ve historically had some of these concepts, things like a supply chain, or things like machine language and machine connectivity inside our own factories. One of the differences is that today we can capture the information in terms of realtime processing. We can capture that much more accurately. We can capture more of it. We are not as limited by, I’ll say, the human intervention in the loop that historically we would’ve had. Now, there are pros and cons to that, but certainly, we have the ability to gather more, and we have the ability to gather it faster.
One of the other things that is important to think about, and we’ve seen this with the pandemic to some degree, is that when you have a weakness, whatever that might be characterized as, in your supply chain, then certainly that presents challenges in terms of foresight to how that may impact operations, how it may impact product availability to customers, or product support for items in the field. And so, the historical supply chain metaphor, we all know that chains are only as good as their weakest link. And so, one of the things that this connected and smart-factory ecosystem gives us is the ability to begin shifting towards a more configurable supply-chain network, as opposed to a fixed and linear supply chain.
Smedley: So, please forgive me if I’m interpreting this wrong. The way I’m hearing then is we are improving our processes, and not necessarily just the technology. So, we need the technology, but we need to have a better process in order to implement the technology. So, it’s a strategy that’s built upon it. So, it’s a layered approach, if I’m understanding where you’re going with this. So that’s where you have smart factories that build on smart manufacturing that enables us to have better workers and better solutions overall. I’m building on all of this a little bit.
Hartman: Yeah, I would absolutely agree with what you said. And you might even take it one step farther, which is essentially a smart ecosystem, meaning certain market segments, certain geographies, maybe certain sectors of an industrial base can also be smart. And in many cases, it has to do with the availability of talent, or the connectedness of the supply network, or the availability of materials, or whatever the value stream might be in that particular instance. But certainly, this is a layered concept to this idea of smart manufacturing, many of the individual technologies that people deploy in order to achieve this have existed for quite a while. The difference being is how they’re deployed, where they’re deployed, and ultimately how they are used in terms of the data that goes in, and the data that comes out, and the ways that people use that for decision-making, and in most cases for optimization.
Smedley: So, this all then says to me, it goes back to now this need of how we have to change the type of worker, where we’ve been talking about skilling and reskilling. But it says to me, we need to understand the type of people that need to be in these factories, that what we need to do is shaping, and encouraging, and doing things differently. So, I guess I want to understand as a professor and someone like yourself, who’s highly educated. You must now need to be thinking that we need different types of people, or we need to educate them differently about the skills they need to have or what we need to be having in those facilities. And COVID, as you just said a few minutes ago, changed or put a spotlight on what we don’t know, and what we do, and where those weaknesses are, and where those strengths are. So, what we can do. Is that correct?
Hartman: Generally, I would agree with that. I think I should be clear. I’m certainly not advocating that we will be running lights out factories across the manufacturing sector any time soon.
Smedley: We want to have that, right? Yeah.
Hartman: I think there’s still going to be humans in the loop in a production sense for quite a while. Now, that being said, I think the work that they will do is more than likely different, and it may not take as many people to do certain work. If you think about the way agriculture years ago got automated, we can farm more land with less people now, arguably speaking. So, there is a certain amount of efficiency that we will gain with smart manufacturing. However, there will still be an infrastructure, probably a much different infrastructure than we might have considered historically. And in many cases in the past, we were automating human labor, the physical part of manufacturing. Well, today we’re now automating, in many cases, the cognitive part or the human part in some cases. And I know that makes people nervous, but I think the way folks can future-proof themselves, if you will, is learning some things about computing, learning some things about networks, learning some things about how to create as well as receive, as well as use certain data and the applications for doing that.
We’re starting to see it, at least here where I am locally, I’ve got three kids in public school and I certainly watch the way they’re being taught certain subjects like math, or science, or even social studies, or language and composition. And much of it is beginning to be geared towards some level of, what I might call, just information literacy, for example. And so, we will need those skills in the factories of the future, because one of the things we will potentially automate away in certain cases is some of the physical labor.
Smedley: But we want to, right? I mean, you just talked about agriculture. We don’t want to do those things that… We want to use our minds versus the repetitive efforts, right? That’s what we’re really talking about? So, if I wanted a career… And the next generation that’s moving in because the Millennials, moving into leadership roles and we know that. And then, the younger generation is going through schools as you’re seeing, as you’re training them. And the baby boomers as we talk about are retiring out. And this is the way the cycle of life works. So, if I wanted to work in a smart factory, and I’m seeing all of these, what things are we talking about, learning more about? Additive manufacturing, quality control, are all those types of things still in play? Or are we saying, everybody’s got to be a data scientist? You got to be a programmer?… There still are key roles that we’ve always had that are still going to critical roles.
Hartman: That’s a really good question, Peggy. I get asked that often. And, even for all of the advances we’ve made, technologically, there are still a role that humans need to do. I still think we will have… Excuse me. I think we will still have engineers and designers. We will still have design engineers, we’ll have manufacturing engineers, we’ll have quality engineers, we’ll have procurement and supply chain specialists, and so on. I think the difference will be, is they will do their jobs differently. They will do their jobs in a way that allows them the visibility and the insight into a much larger contextual environment because of the work of, in some cases, the data scientists or the data analysts, the folks who can ingest information, develop dashboards or algorithms to display it for decision making, and then people who know how to use it.
I don’t believe actually that everyone needs to be a programmer or a data scientist. We will certainly need some of those people. I think, as I just mentioned, even the more traditional rules, people will need to know some basic things about how to interpret data, about how to manipulate data. But I don’t believe that we need to turn everyone into a programmer or a data scientist.
Smedley: When we look at what’s happening right now in manufacturing, we’re starting to see this bigger movement of understanding artificial intelligence, and its role with the IoT, and how we need to apply it, and how we can leverage it to make things better. And we have to start thinking about being more sustainable. We have to think about things… We’re a make-take-waste society. Are we seeing manufacturers understand they’re going to have to step up faster, quicker, easier to be more circular, be less waste, be more resilient?… We talked about lights out. We understand it’s not going to happen overnight, but they have to do it faster than they thought before.
Hartman: I am seeing it. I’m also seeing a struggle that many of them are having in terms of where they might start, how might they do it. Many of them have been down the path on either software or hardware implementations in the past that have not gone very well. We know some of the histories around ERP (enterprise resource planning) systems and MES (manufacturing execution systems), right? And for all of their touted benefits, they’ve become endless black holes that people just pour money into sometimes. And, because we’re talking about computing technologies, as well as other things in this new smart manufacturing environment, I think some manufacturers are a little hesitant. They’re a little bit leery to that jump in with both feet. And so, the challenge is getting them to maybe put their toe in the water first, start somewhere. Don’t have paralysis by analysis, if you will. So, start with a small project, begin to see what value it can add, what benefit it can provide, and then begin to scale from there. To your point about speed, I think it is a challenge for them.
Smedley: If they wait to do that, I mean, we’ve been saying about, put your toe in the water, but right now, it’s a typhoon. If they wait to put their toe that typhoon’s going to take them away. I mean, they can’t wait, just put their toe in. I mean, realistically, their competitors are already making the move and understanding because the next generation that’s making decisions are saying, “We have to do these things.” If they wait and they’re that methodical in their decisions, their decisions are too slow. We’ve been saying with the IoT, “Try it.” And I agree with you, it was ERP, ERP2, MRP2. And, oh my goodness. It was 20 years of, we couldn’t get the industry to agree on standards… But now…If they take that slow decision process, aren’t they going to be the ones truly left behind…They’re going to be left in the dust?
Hartman: Well, I think you raise a really good point and that might be the proverbial catch 22 of the discussion, is that you’re right, they don’t have time to wait. I think one of the differences is, for example, if we use that ERP example as maybe a comparison, is historically, there was not a lot of connectivity between tools, between interfaces, between elements of the organization, to your point, standards evolve slowly, yet because of open source protocols, because the standards community, to be fair, has reacted in a way that’s positive and moving with a little more sense of urgency, there are differences, I think, today. I certainly would not suggest that this technology is plug and play. It certainly isn’t. There’s work that still needs to be done when you implement these technologies.
However, they are closer to that end of the spectrum than they were historically, where you had to write the individual codes, write adapters, write connectors, and so on, for every single piece of technology that had to connect to another piece of technology. So, I believe we are making progress. I think the point I was really trying to make when it comes to putting your toe in the water, is that I don’t think it’s right for manufacturers to wait on a perfect solution. There’s never going to be a perfect solution. And so, it’s important to simply get started. Plan as much as you’re able but get started. And then once you’re started, you can evolve, and react, and pivot as you need to, in order to deploy the technologies or implement the methods that you need to. But I think, the fact that some are refusing to start, because they’re afraid that the next thing might come around the corner, you’re right, the wave is simply going to take them over.
Smedley: And Nate, I agree with you a 100%, because I think they are afraid of what happened. I mean, if you go back, you and I have been doing this a long time, when Fisher Rosemount and Honeywell just fought all the time, no standards and anything… I mean, you got to trust some of these, you’ve got consultants. And a lot of these people are saying, “Look, they don’t understand that this technology is a journey.” It’s not a one and done. And it’s expensive. Right? I mean, that’s what we have to explain to them.
Hartman: Right. I think one of the things that can help is that the technology compared to what it was in the past, right? And again, in many cases, we’d have these enterprise software applications, whether it was PDM (product data management), whether it was MES, whether it was ERP, large hardware implementations, where you might have hundreds of PLCs (programmable logic controllers) on a plant floor. My point being is that those were rather monolithic, in some cases, in terms of their deployments and the way they had to be configured. Today’s technologies are, for better or for worse, a little more modular. And so, companies can begin to deploy certain things without having to deploy an entire enterprisewide platform. In some cases, cloud computing has helped us with that. In other cases, various open-source standards and protocols have helped with that. Or in some cases, just having employees now who have maybe a higher general awareness of certain technologies than they may have had in the past. A lot of the interfaces to these tools look not that different than your cellphone.
And so, I think what we’re seeing here is, the scaling effect is a little easier today than maybe it was historically. But that being said, I think there’s still one nuance we’ve not quite touch in this discussion. And that is the difference between larger OEMs (original-equipment manufacturers) and what are historically called small and medium manufacturers. And I think, therein lies a real difference. And it’s one, I think, that the manufacturing sector at large is going to have to sort out, which is, a lot of the larger multi-division, multinational companies, they have resources, they have people they can put onto these implementations they have, monetary resources. I certainly don’t want to suggest that they can just throw money at it, that’s not the point. But they do have resources where these things can have dedicated teams. Whereas smaller companies often don’t have that scale or that resource level yet, they need to deploy some of the same methods and tools as the large OEMs do. And that presents a particular challenge to them.
Smedley: And this is what I’ve said on this show in the past, if those large manufacturers get lost in the fact that they move so far ahead and they leave the small to mediums one behind they’re going to be in trouble because they need them for a lot of things that they supply in the supply chain. So, there’s a catch 22 there that you’ve said. And I think, that leads me to say, how do you build this smart workforce for smart manufacturing? Because, again, there goes the processes. The people, the process, the technology, it is a supply chain. And if we learned anything in the last couple of years is, it’s a collaborative effort, and you have to collaborate to everyone in the four walls, outside the connectivity, and then the entire ecosystem that we started this conversation. And it’s professors like yourself who get it, that are building the next generation of brilliant workforce, that’s going to lead the next generation, so that we’re not left wondering why we didn’t prepare in the supply chain.
Hartman: Yeah, that’s a really good point. I think, one of the things that is important to think about is that many of the problems we’re faced with do not necessarily have their roots in a particular technology, and they probably don’t have their solution in a particular technology. Much of what we were talking about in terms of the adoption of smart manufacturing, generally speaking, and workforces, and so on, have more to do with how people, individuals interact with each other, as well as how companies interact with each other. In many cases, these are social and economic, or legal and contractual discussions. They’re not necessarily technology discussions.
And so, when we think about the workforce and what we need them to be able to do, having some technical literacy, or technological literacy in some area, yes, that’s important, but we need to be teaching students and exposing them to various problem-solving techniques, various communications techniques, a certain amount of emotional intelligence, in terms of the interaction with people. Various, I’ll say, group interactions, and cultural interactions, and so on, because in many cases, even though we might be talking about an ecosystem potentially geographically, or from a market sector perspective, certainly, globalization is likely here to stay, right? I mean, that’s not going away anytime soon.
Smedley: And that’s the critical part of all this. If we’ve learned anything, we have to learn from, and I don’t want to say mistakes, but we have to learn how we can do it better. I always say, act locally, but think globally. Final thoughts, what would you say? What do we need to be focusing on?
Hartman: Well, one of the things to think about is, going back to the example we talked about before, with respect to the automation of things and the accessibility of certain technologies, depending upon the types of products someone might want to make, we may not… And to be fair, I don’t have literature references or historical references necessarily. This is a little bit of anecdotal, but I think, it gives people something to think about. If you think about the capital infrastructure that was necessary to create a manufacturing plant historically, those numbers were quite large. If you think about the monetary outlay that had to be there to hire a workforce, to train a workforce, and so on, those numbers are pretty large.
Hartman: Today, you can use either low cost or open-source 3D design tools in a cloud-based environment to either access manufacturing. And by that, I mean, production capacity, by way of the cloud, or you can build your own three-dimensional printer for a few hundred dollars with parts you can buy online. And you can begin making things. Now admittedly, there’s probably very limited market for those things. But my point is this, the barrier to entry to get into manufacturing, so to speak, I’m not sure has ever been lower, outside of people making their own, say, furniture or tools from a subsistence living point of view. So, if we think about our ability to design and make things, smart manufacturing has enabled a lot of the work that we’ve done in the past that required groups of people to do, to become rather individual. And so, that might be a topic for a different show, but that’s something that I think is important in this smart manufacturing environment, is that things have the ability to become more individualized.