In advance of a May 20 webinar, Roadmap to the Autonomous Plant, I had the opportunity to speak with Joe Perino, Principal Analyst with LNS Research. Joe will join Emerson’s Stuart Harris, Group President Digital Transformation and Mart Berutti, Vice President, Digital Transformation.
In this podcast, Joe shared with me elements from some of his recent research, On the Road to Industrial Transformation (IX): Remote Operations & the Autonomous Plant. He’ll explore his findings in greater depth during the webinar.
Have a listen to the podcast or read the transcript below and don’t forget to register for this autonomous plant webinar. We’ll see you there!
Podcast: Play in new window | Download (Duration: 26:42 — 20.9MB) | Embed
Subscribe: Apple Podcasts | Spotify | Email | RSS | More
Transcript
Jim: Hello, everyone. I’m Jim Cahill and welcome to another Emerson Automation Experts podcast. The advancements in sensing, computing, and communications technologies have enabled manufacturers to rethink how they do business to operate more safely, sustainably, reliably, and productively. And the COVID-19 global pandemic has accelerated these digital transformation efforts.
Two areas of focus in these digital transformation initiatives are remote operations and autonomous operations. Today, I’m lucky to be joined by LNS Research’s Joe Perino to preview some recently published research in this area to be shared in a May 20 Autonomous Operations webinar, and you can visit emr.sn/autonomous to register for the webinar. Joe is a principal analyst at LNS research with extensive expertise in asset performance management or APM 4.0 for the process and asset-intensive industries and infrastructure.
He focuses on emerging technologies, including the Industrial Internet of Things or IIoT, big data, cloud, advanced analytics, edge computing, digital twins, robotic process automation, blockchain, connected frontline worker, operational excellence, remote operations, and the autonomous plant in the process and hybrid industries.
Whew… that’s a lot! Welcome, Joe.
Joe: Thank you, Jim. Glad to be aboard today.
Jim: Well, it’s great to have you. So, why don’t we just dive into this? And maybe let’s start with a very basic question to ground our listeners in the discussion. What are remote and autonomous operations? And is there a difference?
Joe: Well, there certainly is, Jim. So, we’ve had remote operations, which is really the ability to do remotely what a local board operator would do at a plant be that in a control room, or simply engaging with the control system at that facility. And that’s gone on for a long time. Industrial gases, water and waste, probably started in drilling in upstream. We’ve seen it for probably 15 years now or more on offshore platforms in the Gulf of Mexico, for example, where there’s actually a control room onshore, which can backup or take over if the operators are tied up, or there’s an emergency or some other situation on the platform. So, this has gone on for some time. Now, autonomous operations are different. And of course, it depends on your definition of autonomy. And one of the things that popped up in our research is that most people still think autonomy are programmed operations within defined boundaries, which is sort of basically a way of saying automated. And of course, autonomous will require a lot of automation.
But autonomous operations are really those that can react or respond to unexpected events during operation. In other words, they’re not programmed, they can adjust to things that they weren’t programmed for on the fly. And that’s a lot different. And so, it’s gonna require automation, but it’s a lot more than that. And so, that’s the basic difference.
Jim: That makes sense. And in your research, you’ll be expanding upon in this May 20 webinar. You highlight eight key motivators for the move towards autonomous operations. Can you highlight a few of them?
Joe: Well, certainly. So, I listed eight and I’ll take a couple. Number two, I said was to better manage risk. This is really important. And I would say that the average process industry plant, particularly the large continuous ones don’t have a view to their operational risks that they would like to have, which also includes the human performance factor as well as the mechanical, physical process factors that are going on in the facility. And so, one could argue that with autonomous systems, which can detect and respond much quicker and be able to absorb the massive volume of data and make sense of it, that in fact, with that kind of advisor alongside or actually steering the plant, that you could argue that the plant will actually operate more safely, and therefore has a lower risk. Now, that’s balanced off by the fact that you have to trust these systems. Now, we placed a great deal of trust in safety shutdown systems today. And in general distributed control systems, these have evolved a long way, they’re very, very good now. But it wasn’t always the case.
And so, risk is really important to them. Of course, safety is important. Another key point of this is that it’s gonna require upskilling of people. Because as you move toward both remote but particularly autonomous and the person who was the board operator isn’t gonna be needed, certainly as frequently, then what does that board operator do? Well, you certainly don’t want to lose that talent. So, they become more managers of the technology making sure the system is running correctly, than they are actually interacting with it and steering it. And so, that’s gonna require some upskilling. The other thing, of course, is that you’ll probably have fewer people overall at the facility. So, those who are there may need to learn multiple skillsets, become Swiss knives, so to speak. So, that’s an opportunity to upskill people. And that’s gonna be a real challenge for most companies. This as you know, in the process industry we don’t exactly have a long line of people trying to get in and work at the refinery these days. So, those are two points that are really important.
And then the last one I’ll touch on is that if we can use these systems to detect deviation from an unwanted condition, or simply from the plan, and correct them factor, and then recover from anything that also is deep, because the deviation for an upset or an incident, then we actually improve our resiliency in the facility. So, if you have lower risk, you’re more resilient able to steer through the potholes and bumps in the road, that helps you actually become also more sustainable. So, those are the three or four that I wanted to highlight.
Jim: Well, and that is exactly right. People’s jobs will have to change as technology is taking on more of it, the care and feeding of the technology, the broader skill set to be that Swiss Army knife to deal with things as they come up. And with all these predictive types of technologies, to be able to make sense of that and do things before it becomes a problem. So, yeah, I could see that and can’t wait to hear what some of the other ones are there. So, what industries have made the most progress in advancing towards autonomous operations? And what are their key business drivers that’s taking them down this path?
Joe: Well, first, let me contrast the fact that discrete industries are in general, but it varies widely, more advanced at using autonomy than the process industries are. So, I’ll give you two examples there. One is in automotive assembly, where robots do a great deal of basically assembling the chassis. And that’s all pre-programmed, and things move through and people don’t touch it. And the other one is in electronics and semiconductors, chip plants. Just the other day, I was watching television, and they were talking about Intel’s investment in the U.S. And they had a person that had to go down and get in the suit and clean suit to be able to go into the plant. And it’s extremely highly automated these facilities, and they also use robots, of course, a lot. And so, those two are really good examples of close to autonomous facilities for large portions of their operations. Now, when we swing over to the process industries, and I mentioned some of these things before, you know, there’s a mix of this that’s gone on for quite a few years and I mentioned some of that in the answer to your first question.
But the key business drivers for a lot of these things are the same business drivers that we have today. People want to lower cost, increase revenue, push throughput through, operate safely, and so forth. And I would say the best examples of that probably have been found in the mining industry, at the mine, and not necessarily at the ore processing facility. So, when you look at people like BHP, Rio Tinto, they’ve made huge amounts of progress in the space toward autonomy. The vehicles, the shovels, the trucks, all are basically being fairly autonomous, there is a remote operation center of course, it’s watching these people. And they’re actually able to move the raw ore to railcar, and go from the interior of Western Australia all the way to the coast and then to be loaded on and export it if it’s a case of iron ore or if it’s coal, for example. So, there’s really good examples, of course, of that. And then the other ones I like to mention that are in the news today, you and I’ve been watching this and witnessing this are the autonomous landing craft on Mars, the Perseverance craft and the autonomous helicopter that goes along with it. That’s an impressive feat.
And then I’ll mention one more thing. IBM actually has built a trimaran boat, that they are gonna sail autonomously, across from Portsmouth, England to Portsmouth, Massachusetts. And that should happen at the end of this month, or sometime in June. So, with every step, we’re increasing our ability to do these things autonomously. Again, still, though, with a remote operation center loosely, in the case of the spacecraft, or more wired through the case of networking to that facility, or vehicle, or machine.
Jim: I think it’s fascinating to see every day you see more and more examples, the technology has just blossomed to enable people to be able to do that and have that remote look in. I saw even F1 Formula racing with how much is computer and people in the back kind of helping the driver of the car out get through the race. So, it’s just across all the industries.
Joe: Oh yes, Jim. That’s an excellent example. People don’t realize how instrumented the vehicles are these days in that sense. But I’ll give you one example where we’ve been led to believe that we have are further along in autonomy than we are and that’s with autonomous cars. And we’ve all heard in the news that some of these vehicles are just not as autonomous as we thought they would be. In other words, you still need the driver behind the wheel. That’s come up with Tesla in incidents recently. And then there’s actually been a response to that, in fact, Cadillac vehicle has a camera on the steering wheel. So, if they don’t see the face of the driver there, then you will not run autonomously, you won’t take over. So, there’s ways around this kind of thing. But I think we’re still quite a ways away from letting cars go down the road at high speeds without a driver in the front seat.
Jim: I may be a bit old school there, but I still want to keep my hand on the wheel, I’m not ready to quite give all that up. Now, in the recommendation section of the report, you describe a system of systems concept. Can you share an overview of this concept with our listeners?
Joe: Absolutely. First of all, a lot of the things we think are systems today. And I will show a picture of what I call the performance maximization side of the facility or the gas pedal, I also call that the utilization side. A lot of us think that these applications which have a cascade of data connecting them, or a system with the control system being at the bottom of that stack, but in fact, the only thing that’s really a system there is the control system. And the rest of the applications are actually being integrated and managed as a system by humans. And that’s just one side. The other side, if you can imagine if the listener can imagine two large gears, one gear is that utilization gear on the right-hand side, that’s that performance one. It does the throughput, the yield, the energy management, all of that stuff. The other wheel is the availability gear, which is your maintenance inspection, reliability, integrity side. And so, those are the two big wheels that must mesh together. You can’t push utilization if you don’t have availability. So, one enables the other.
Now, the problem that some people face today, refining is a good example, is that the demand is down so that you don’t need maximum capacity anymore. Well, if you don’t need maximum capacity, then you only need as much availability as is necessary for the demand to use the utilization that’s there. And that’s one of the reasons why I think we’ve been in a bit of a pause in terms of progress around asset performance management and things that enhance availability because of the situation with the pandemic causing a drop in demand, and also the oil price situation which was independent of that prior to that. But we expect that to pick up now as we get out of the pandemic and so we should see demand increase. Now, on the flip side, consumer product goods, food and beverage and particularly pharma, the demand there has been high all the way through this. And those companies particularly paper company like Georgia-Pacific or International Paper they’ve been putting out, they’ve been running at max, and so they’ve been actually wanting to squeeze more availability out on these two sides.
So, that sets up the two gears in the middle. So, the question is what actually is gonna manage this? Each one of those needs to be systematized, each one of those gears. But what’s gonna manage those and make them mesh together and optimize how they work together? Well, right now, these are humans that are doing all this, not a separate system. So, when we think system of systems, there’s gonna have to be something that plan schedules, and optimize and manages these two big gears together in the middle. And well, I don’t have a picture to show during the podcast, you’ll see this at the webinar. And then on top of that, you’re gonna have other smaller gears like procurement, inventory, and other functions that basically are driven by these two big gears in the middle, and then you’re gonna have a remote-control center attached to that. So, you can almost imagine that picture with the big gears, it’ll have small gears around it and some kind of a brain or workflow orchestration with some sort of optimization that is managing this together.
This is what we’re getting to or trying to drive through with autonomy and, of course, a lot of automation in that. But we have a long ways to go to make that happen. So, that’s the picture I’m trying to paint.
Jim: Well, it definitely sounds like with the market drivers such a big factor that we can expect to see it moving along different pace in these different industries there and really driven by, I guess, if those two gears are matched and going well, then the lot more advancement made there. With all these things going on here, what are some of the foundational enabling technologies driving us towards autonomy?
Joe: Well, one of the biggest things, of course, is to have good data, and have access to good data. And that’s not only true at the plant level, but it’s true when you need the edge in the cloud to provide the scale that you need, which is also another reason. And so, one of the big challenges today is how do we create a scalable enterprise architecture that can provide access to data at all levels, and condition that data and put that data in a context. And I will use the word metadata structure, so that an application like a planning application, or an advanced control application, or some type of digital twin monitoring of pump, a compressor, etc., can access that data consistently, and then turn around and give the best answer, and then have that flow back down to the plant. Along with that is that we’re going to have to design an ecosystem architecture that isn’t the layered air-gapped data interfaced structure of the original ISA95, or the Purdue model. These things are being compressed and collapsed down so that basically, layer three is being compressed down to be very close to layer two and parts of layer four as well.
And that means that it’s gonna be more challenging and more complex, to design that architecture and also to secure that architecture. And I’m talking also about cybersecurity, they’re very, very important because now you have the ability to take a sensor to the cloud and not even go through the control system. And a lot of people will be using a wrap and extend approach around existing systems to extend their life and add capabilities until we have more open systems that basically then, you know, when your basic systems that are at end of life, and you have to replace them, you will probably replace them with newer systems that have more capabilities, and are easily more integrated and more open, whether that’s OPC UA or whatever other method to be able to do that. So, foundationally access to data in and enterprise architecture are gonna have to be there. And of course, for many people, depending on how far away you are from a facility with remote Ops, you’re gonna have to have bandwidth and that’s gonna have to be very reliable. So, I’ll leave those three or four things there as foundational elements to be able to do this.
Jim: Well, that’s a good view on the technology side. But what about what’s required to make the leap from automated to autonomous and deliver value that’s outside of the technology piece?
Joe: So, we touched a little bit on this a little bit earlier. This is the organization and people side. So, it’s not only upskilling people but what’s your organization gonna look like when it is thinner and flatter at the plant. And you have to actually delegate more authority to those people at the plant because they’re either going to be driven by the systems, or they’re gonna be managing the systems and what does that look like? So, for example, let’s say that a technician is out and there and he’s inspecting a pump and says, oh, we need two seals for this pump and I’m actually able, with my tablet to check the stock, and I see we only have one seal. So, I want to press a button that automatically orders that other seal. I don’t need all these other approvals, which right now is typically, the technician would put a note in there, then you’d go back to the planner and scheduler and they would update that, and then they would initiate a purchase order request that’s in the EAM [Enterprise Asset Management] system. And then that would flow to the ERP [Enterprise Resource Planning] system, might be the same in the case of SAP or Oracle, for example.
And then that part is ordered and then the part is, there’s a purchase order that’s issued, and then the part is ordered and finally the part comes into the plant and the plants checked in. That whole thing should be automated. Because we know what the reliability of that pump is expected to be, we know what spare parts need to be on hand, we already know what vendors we’re gonna buy from. That whole thing can be automated and made autonomous so that the technician working with the physical equipment, the cyber-physical system automatically has all that support to make it happen for them. And then when in fact, the whole thing will end up being scheduled. And then the technician some weeks ahead, or whenever the part comes in, takes the part. He’s already had a schedule; he goes out and he fixes the pump. So, that’s just an example of these kinds of, I’ll call them cross-functional workflows that need to be looked at, streamlined, and then automated and then made autonomous.
Jim: I can see really squeezing out that extra time in there particularly all these sensors and predictive technologies that in advance could trigger that process. So by the time the whatever asset gets bad enough condition you already have what you need to replace it and everything. You can see that coming together over time.
Joe: Let me add one more thing to that, then you have to ask yourself the question, do I really need all these other support groups sitting at the plant? For example, engineering, if they have access to the information? And the answer is that you, unless you’re doing a startup, the process engineer, who’s a unit engineer, for example, as an example, can simply be remote. Maybe they’re there two days of the week, and they’re remote three days a week, because why? Because they have a data pipe into the plant, and they could do their job remotely. And so, this, you have to look at the organizational structure of the facility. And actually, this will have repercussions up even to the business model how you want to interact not only with the plant, but with the rest of the supply chain and your customers and suppliers. So, all that is fair game.
Jim: That sounds like a digital transformation in action. So, let me finally, last question, I’ll throw your way. And I don’t want to steal the thunder from the webinar. But where are some of the leaders in autonomous operations investing now? And are they taking an incremental or an all or nothing approach to it?
Joe: Well, I would say the leaders are investing in the initial infrastructure that you need that I just mentioned around data, and around bandwidth, and around the call at the manufacturing Data Cloud, the data lake. They’re all figuring that out now before they put on all these other apps, because you can’t scale the application. A lot of people have done testing in technology and they say, okay, that’s work, that’s great. How do we put that across 100 more assets at the same plant, or across 1,000, or 2,000 assets across 30 plants in 30 countries? You know, imagine if you’re the size of BASF with 12 divisions all over the world, and a lot of other manufacturers the same way. You know, whether you’re a Stanley Black and Decker or you’re Owens Corning, where you’re making shingles, and other insulation that goes all over the place. These are firms with big footprints. And so, they need to figure out how you’re going to scale all that before they just go off and basically dress up the nicest plant that they have, which is often what we call the lighthouse example, where you pick your best plant, you have the technology and then you say why can’t I scale that everywhere else?
And that’s usually because the other plants aren’t ready for it. The people aren’t let alone the foundational technologies from out there. I would say then that people are taking a more incremental approach rather than an all or nothing approach. But the incremental approach is being taken across several fronts. Our data show leaders are doing many things all at once, but they’re not going for the home run, they’re going for base hits. And I’ll say this, when you look back at the data we have on autonomy, you have to realize that even in a facility, some areas of the facility cannot progress without other areas of the plant progressing with them. So, even though you’ll see a set of data that shows where people are in the progression, what we can’t show is the cross-functional connection between that. And then that also means that every company is going to have to decide how autonomous they want to become, what they’re comfortable with, with their technology with their people with their organization.
So, eventually, you’ll get down to where the limitation is around their culture, and their appetite for risk, and their trust in their new organization and the technology.
Jim: And probably their market conditions when they’re trying to figure out their path forward, there’s a lot of variables in there. But the big thing is technology continues to advance so the capabilities to do it. And it’s just finding the culture, the leadership, the risk, and everything else to drive your path forward.
Joe: That’s right, Jim. I have no doubt that technologies will mature but change management will always be the big challenge.
Jim: Oh, yeah, yeah, the people side of things. Joe, we really appreciate you joining us today and giving our listeners some insights into the autonomous plant and remote operations research that you’ll be sharing in the May 20th webinar. And for everyone listening again, that’s emr.sn/autonomous to register. Thanks again, Joe.
Joe: Thank you, Jim, for having me.
End of Transcript
