Moving Beyond Automation in Process Industries – Transcript
The process industries are advanced in many ways compared to traditionally discrete industries, but that does not mean there is nothing to learn. In the episode of the Future Ready Podcast from Siemens, we sat down with Mark Hindsbo and Rebecca Vangenechten of Siemens to figure out what the next steps are for process businesses beyond automating their lines.
Nick Finberg
00:09
Hello and welcome to the Future Ready Podcast from Siemens. I’m your moderator, Nick Finberg. Today I’m joined by our co-host, Mark Hindsbo, the Head of Operations Software at Siemens Digital Industries
Mark Hindsbo
00:21
Great to be back on again together. I really enjoyed our last conversation.
Nick Finberg
00:25
Yeah, it was really fun. I’m glad to have you back again, Mark. And today I want to talk with you about the process industry more specifically. And to do that, we have Rebecca Vangenechten, the Head of Automation and Engineering Systems for Process Automation. It’s great to have you, Rebecca.
Rebecca Vangenechten
00:42
Thanks for having me. Happy to be here.
Nick Finberg
00:44
To start off today, let’s set our focus on maybe what’s already available today in the industry and what’s up next for transformation. But maybe let’s start with the foundation in automation.
Mark Hindsbo
00:58
So in some ways it’s a really, really good thing that we have the automation foundation layer in place. You know, we’re building on top of that. I think it was you know Isaac Newton who said if I’ve seen further, it’s because I’ve you know stood on the shoulder of giants. And I feel we stand on the shoulders of giants of the people that put the you know, foundational automation layer in over the last it number of years. But there are two things I think we need to do with that automation layer. One it is in a lot of times implemented as a fixed layer, as static if you want, automation. And what our customers are really looking for is to be able to more dynamically adapt their production. So, we need an automation layer that can also be adaptable, that can change the processes and not be fixed.
And the other thing that we also need to take a look at if that might be a problem is you know as engineers, we have this tendency of if it’s not broken don’t fix it, or if I build a new factory, whatever automation I put in is then frozen in concrete at that point in time. If we want to have more dynamic production, we need to be in a place where all of our production environments are continually updated and continually kept in sync. Otherwise, we won’t be able to you know keep this network of new adaptable technologies in sync. So, I’d say the automation layer itself needs to be adaptable and then we need to make sure that we’re building factories and process plants where the software layer is continually being updated.
Nick Finberg
02:30
I can wrap my head around a lot of what Mark was saying, but I know I’m only thinking about it in regard to discrete manufacturing. So, for myself and anyone else kind of more comfortable with vehicles rather than bioreactors or mixing machines, could you talk about what this distinction is that we make? Is it a real distinction? Is it some kind of way to think about the types of things that would go into a solution for the process industries?
Rebecca Vangenechten
02:58
Well, there’s definitely, of course, differences when we think about automation in the process industry that has to do with the fact that process industry they take high value around safety, security, and yes, there is an element of continuity in some of the processes that are being run specifically when they’re continuous. But we also see here a lot of the things that Mark was just mentioning about dynamics and adaptability and flexibility that is more and more required, more than ever before, I would say, also in the process industries, which is a very quickly changing environment. So, I would say the paradox that we see today in the process industry and with leaders in the process industry is that there is already a high level of automation, but with that high level of automation alone we’re not necessarily getting a high level of adaptability or flexibility. So that’s really the question I guess that we’re trying to answer together here today.
Nick Finberg
03:55
For sure. Building on that, if automation is already well introduced in the process industry, what’s holding companies back from becoming more adaptable?
Rebecca Vangenechten
04:04
Well, I think it’s a, as always, right, it’s a combination of things, yeah. It’s more flexibility that doesn’t jeopardize necessarily we’re still running our processes in a safe manner that doesn’t jeopardize on quality and the trade-off of having more flexibility, but does that then automatically mean that I have less efficiency. So how can technology really break that paradigm and bring that adaptability, at the same time running processes, extremely reliable, yes, with a level of efficiency, yes, ensuring quality of an end product. That is really what up until now in my opinion has held the industry a bit back from embracing this resilience that we will be needing more and more. Yeah. And in the past, I think that the process industry sometimes, you know, made it life a bit easy and simple.
So, you know there’s lots of regulations. around safety, around product quality and GMP, for example, in in in the food and beverage industry or in the pharmaceutical industry, but they were to some extent good excuses that I think today we’re seeing that, you know, regulatory bodies are helping us to overcome them. Now if I take the example of the pharmaceutical industry, an industry that I have my background in where quality by design principles in the meantime have been introduced for over a decade now and we see that when producers can demonstrate process know-how that they will be allowed to operate within the so-called design space that allows for more process control variation. So, process understanding is prevailing over process rigidity and that’s really something where I think technology then also steps in and supports that.
Nick Finberg
05:56
Maybe linking this back to some of our previous conversations, Mark, how does operations software kind of help with this goal of insight that Rebecca’s talking about?
Mark Hindsbo
06:07
I think operations software actually helps in three different ways. You started with insights. So certainly, we have a level of analytics built in that allows you to analyze some of these complex relationships because sometimes your root causes are you know can be tough to find or maybe even counterintuitive. You know, you might think that it’s a specific machine that that you have an issue with or part of your processing. but it could in fact be a specific supplier, as Rebecca said, that had a slightly different quality of some ingredient that went into the process. So being able to trace all of those different qualities across your DCS system, your MES system, your supply chain system and so on can be pretty tricky. So yeah, we do have a set of analytics that can allow you to track those root causes and figure out how you can optimize your production. But then I think there are two other ways that are also important. Customers are asking for flexibility. It is an ability to produce different things or even individualized things that requires you to have a very, very good handle on your recipe management. and changes in recipe have to propagate into the factory very, very quickly.
And they might have to propagate into different factories. So, you might have a factory in Brazil with one humidity and one factory set up and you might have another one in northern Europe, right, in a different climate and so on. How do you manage propagations from changing your upstream formula to changing it with quality, as Rebecca said, downstream? That’s something where our op software also helps because it has the connectivity into your product definition system or your PLM system. So, you can very quickly go from changes in in in formulation or recipe into changes in production and you can steer that with quality across a network of multiple different factories. And then last but not least, we talked about automation. We have a very, very clear link to the automation layer. So propagating changes in process down into the automation layer also. is in enable and one of maybe one of the things that we uniquely can deliver with Siemens where we have both the hardware and the software stack coming together
Nick Finberg
08:27
Rebecca, could you talk a little bit about how companies or how we’re working with people to kind of deploy these software-defined systems for the process industries?
Rebecca Vangenechten
08:35
Well, I think it it’s about making automation behave more like modern software at the end of the day and once engineers and operators have embraced that idea it’s quite attractive let’s say to convince them of how that can help us with many of the examples that that Mark was already giving to really achieve the flexibility without not respecting what makes the OT space the OT space. Now I think that’s always important then to keep in mind is that when we talk a lot about how automation can we should integrate IT-like principles, but what does that mean for security and safety of the OT space? We still need to maintain that high availability in particular for the processing industries. We still have secure networks, data connectivity. And then the other element that we haven’t talked so much about, but what is very particular for the process industries is the life cycle of such a place. the long-term serviceability of these systems that in a way is not something that IT systems are known to be extremely fantastic in, so to say.
Nick Finberg
09:48
You’re more likely to replace a server after a couple of years rather than try to fix it or maintain it. Exactly. Do either of you have any like concrete examples of maybe what this looks like for some of our customers? I know names can be tricky at times, but in some instances.
Mark Hindsbo
10:03
I can jump in and say, hey, we actually had two, both Rebecca and I just came from Hannover Messe, and I think we have two great examples there with names that illustrate some of this. So, one was we showed a collaboration we’d done with Pepsi for their Gatorade production for the upcoming World Championships in soccer or football depending on yeah where you are in the world the thing that is actually played with the foot no and they have both their own factories they then have these pop-up factories and they have subcontractors they work with. And depending on who gets to which stadium, it kind of ironically, you know, if I’m wearing a red t-shirt today, if that was the color of whatever team I’m supporting the red Gatorade might all of a sudden become more sought after in a specific geographic market. And we showed a system where they can on-demand see who has the right ingredients to produce red Gatorade. How can we dynamically dial up this pop-up factory or dial up the base load in our other factory?
Maybe the equipment that exists in one factory is slightly different than in another and therefore the recipes that need to be sent are slightly different or have different, you know, set points and so on. So, we actually demonstrated how this complex supply and demand that comes in can dynamically be distributed in an in a network of factories of which some are permanent. Some are pop-up and might only exist for this specific week. And some of them are subcontractors and you can distribute the demand, and you can in real time figure out what the ingredients are, and you can get your recipes out. So that’s one example of it being real today, this sort of dynamic planning, changing of recipes, changing of different products that get produced.
Rebecca Vangenechten
11:55
And that very much resonates with what we see. I mean that was an example of food and beverage, but also in again fine chemicals, biotech, pharmaceuticals, we see that industries there are moving away from these large monolithical plants to much, much more modular and then to Mark’s point also more software defined plant concepts that at the end of the day of course have then an evolution from a structural perspective, from an architectural perspective, but also from a cultural perspective. As was said before, we’re moving away from that principle of don’t touch a running system towards a much more continuous in a controlled environment improvement along the way of all these layers yeah.
Mark Hindsbo
12:44
And I’ll give another example that really resonated with me in a in a different industry you know so BASF and Accenture had built this digital twin of their laboratories which are small factories super small factories right two or three pieces of equipment each, but they have 33,000 of them and usually they only operate while the scientist is at work, right? And the scientist will be testing out a new formulation. So, the processes are very changeable, very adaptable. Oh, maybe I need to do this. But they’ve built digital twins of each of these, you know, and they are very different, right? You might have five pieces of equipment in one laboratory; you have seven in another. right but they can quickly adapt the digital twin. The digital twin can see what the scientist is doing. The digital twin of the factory then trains the small robot to do what the scientist did and then when the scientist go goes home the robot can continue the experiment overnight and you could even say, hey, deploy this thing in one of my other laboratories to drive it. And then they get a lot more efficiency of these things that might only have been 30% utilized before, right? But it’s a very heterogeneous environment, it’s a very dynamic environment and so on. But by building this digital twin and having physical AI in the loop that can be trained on demand, then they can also manage very disparate processes.
Nick Finberg
14:13
Oh, that’s super interesting. Kind of like donating compute resources when your computer’s offline. You’re donating the capacity of production.
Mark Hindsbo
14:21
No, but you can you can imagine this lab space, right, where a scientist goes to a you know place where the chemicals gets get spun and then it goes over and puts it in a different little mixer and then there’s probably a small oven that heats something or whatever it is and as they’re experimenting the recipe changes dynamically. So having a digital twin that they can keep up to speed with that. and then move it into sort of continuous production is an interesting concept.
Nick Finberg
14:51
Yeah, that is so cool. thank you both so much for sitting down. This has been a great introduction, at least for me, on what is happening in the process industries right now. There are still a lot of questions I have for the both of you, but I think we’ll hold on to them for our next episode. Thanks to the audience for tuning in to another episode of the Future Ready Podcast from Siemens. We’ll be back soon with Mark and Rebecca diving into distributed production. But until then, be sure to subscribe and check out our other amazing episodes on AI, pharmaceuticals or maybe some of the semiconductor industry. We hope to have you back soon.
