How to unleash the power of data through machine sensorization and digitalization

Since 2011, these developments have been referred to collectively as Industry 4.0. Machines have the ability to communicate with one another autonomously, thereby optimizing process flows. Industry 4.0 clearly relates to networking in the industrial field, and sensors play a very important role in the value creation chain in this area.

The prerequisite for communication is an abundance of information, which is what smart sensors deliver. The smart factory is a prerequisite for Industry 4.0. Every sensor, every machine, and every human involved can communicate with and among one another at any time. This information exchange does not end at the factory gates, however.

The interplay of edge and cloud also allows production and data management from and to the outside. This intensive cooperation between technology and humans makes the process more transparent, productive, and profitable. Divya Prakash is Director of Business Consulting, Industry 4.0 at SICK Sensor Intelligence. He has more than 30 years of experience in the manufacturing industry including leading customers in their digital transformation journeys.

Prakash is well-versed in emerging IT and OT standards, edge computing and cloud implementation. And he has experience in implementing automation and information solutions across enterprises. Welcome, DP.

Divya: Thank you, Mike. It's my pleasure to join you this morning.

Mike: We are so happy to have you here. I'm sure this is going to be a very interesting conversation. So let's get started. Digitalization has made machine data available for remote diagnostics and analytics, as I mentioned previously. How has this advanced the promise of productivity as a service, equipment as a service, or maintenance as a service? How has that availability of data enabled this?

Divya: All of those that you mentioned, productivity as a service, equipment as a service, maintenance as a service, are all offshoots so far, as we're moving forward with the digitalization. So before we move forward, it's always good to see where we're coming from, right? So digitalizing of the sensors, digitalizing of the production data from paper system into Excel spreadsheets, we’re all aware of how it has now given us access to a lot of raw data coming in.

The next step is to combine this raw data with some contextual information to make them into actionable data. There's no shortage of raw data. So we can then use this actionable data in our daily operations, and this is what we are now calling digital transformation. How do we transform our operation and our businesses using this digital data?

And digital transformation is not new. It is happening all around us, as we move forward to improve our existing processes that help diagnostics and analytics. And, when I say it's happening around us, it's happening in our lives too. If we look around, the way we listen to music today, the way we use smart cameras at our front doors, it’s digitalization, a digital transformation, and.manufacturing is no exception.

Internet connectivity has made it much easier for us to connect to remote installations and perform all the services you mentioned earlier. So the availability of this data, whether on-premises or in a secure cloud location, opens a lot of opportunities for suppliers, OEMs and machine builders.

Traditionally, once the machine builder sets the equipment and successfully installs it, they have very limited insight into their equipment. Imagine having access to real-time operational data to see how well your equipment is performing, analyzing if the machine has been used ideally, for example, within the design specifications, and performing remote maintenance where you can see metrics drifting or components in the field, this enables the machine builders to provide additional services to their customers, like maintenance as a service.

So, digitalization is opening up everything that you mentioned, productivity as a service, equipment as a service, maintenance as a service, access to data so you can do your own R&D. That's a lot of things that digitalization is making changes into the industry.

Mike: What about the actual technologies?  You talked a little bit about the remote connectivity and the smart sensors, the ability to collect that data, whether it's doing something with it on-premises, at the edge, in the cloud, but can you talk a little bit about the specific technologies that have been the most important in enabling this digitalization and the remote connectivity?

Divya: Absolutely. Industry 4.0 or the Fourth Industrial Revolution, with its underlying technological advancement, has made this possible. The Internet is pervasive now. Pretty much it is coming into place, and it is one of the underlying technologies that is helping us. The Fourth Industrial Revolution, with its digitalization and networking on machines, has been changing our life for some time now. These new technologies have allowed the physical and virtual worlds in production logistics to merge, right? And you mentioned that in the introduction, the new cyber-physical system, but it's not just on-premises, now it's connected to the cloud. So these developments since 2011, these developments had been collectively referred to as Industry 4.0.

Machines can communicate with each other autonomously. They were optimizing process flows. Industry 4.0 clearly relates to networking in the field. Now, sensors play a very important role. So when you're talking about technologies, the most important underlying is it gives the ability for machines to talk to each other, to see each other and communicate. So this communication is an abundance of information, which is what you're getting from smart sensors.

The other bigger items like Industrial Internet of Things, you hear these buzzwords out there, big data analysis, machine learning, artificial intelligence, edge computing, cloud computing, these are all helping us get a better insight into machine performance as we move forward, as well as moving from, as we call it, predictive maintenance to prescriptive maintenance. So that is what is happening with this underlying technological advancement.

Mike: Right. If you have no digitalization, you have no data. You need a sensor to collect the data in the first place.

Divya: That's where the rubber meets the road.

Mike: You touched on predictive and prescriptive maintenance, and that seems to be one of the low hanging fruit for a lot of the digitalization initiatives, where it's simple to explain to someone, "Here's an application where you can reduce your downtime; you can increase productivity." But besides the maintenance-related applications, what other applications are benefiting from this digitalization? I mean, is there a use case that you're seeing for acceptance testing or virtual commissioning? And how many manufacturers are actually looking for things like those productivity gains or even something like flexible engineering capacity?

Divya: Good question. I mean, this machine digitalization is making access to information, right? So the most important thing is sensors are very important, having sensors to the machines to sense the process variables, and make them available for either machine control or to perform production counts, quality control, besides productive maintenance. There's a lot of different applications that you can do.

So these same sensors are also transmitting the data to higher-level systems for data analysis or to feed into digital twins. We have cases where machines have been commissioned with a very skeletal crew at site, while the majority of the team watched and monitored the operations remotely, because now everything is connected to the facility and you can now watch remotely.

During this pandemic time and just lately, a lot of customers, because of the travel restrictions, have done virtual factory acceptance testing where they could walk in through a camera, but also monitor all the data coming directly on a dashboard coming in and actually see the operations. So the possibilities are endless of what you can do. Once the data is available, what you do with the data, and how you use the data, there is no limit to how you want to use it.

The same data can be used differently by the quality people, looking from a quality window, people who are in operations, people who are in maintenance, because they all look at the same data differently, but underlying data coming from the sensors gives you a lot better understanding of uptime, downtime, all the other aspects. So there are different applications.

Mike: I suspect we've just barely scratched the surface in terms of some of the applications that will be available with this inordinate amount of data that we have available to us now. There are probably things we haven't even thought of that 10 years from now will just be commonplace, because we have the access.

You had talked just a minute ago actually about the digital twin, and that was the first thing that popped into my mind when you were talking about the remote connectivity, and the ability to have your machine installed somewhere, and you can watch it, and see what it's doing, and look at its operation. To a lot of people, that term "digital twin" is something they hear a lot, but they're not really sure what it is. So, can you explain how machine digitalization can be used to create a digital twin and what the significant advantages of that might be? Actually, can you define the digital twin for our listeners who might not be familiar with it?

Divya: A digital twin really is, in my understanding, a virtual model designed to accurately reflect a physical machine or an asset. It's trying to make it virtual, a model of the real machine. So this machine being virtualized is outfitted with various sensors related to vital areas of functionality. So the most important thing is the physical asset needs to have a lot of sensors that pick up the data so you get a real-time view of what's going on.

These sensors produce data about different aspects of the machine's performance such as energy output, production counts, temperature, vibration and more. This data is then relayed to a cloud system and applied to the digital copy. That digital copy is what we call the digital twin. You're getting the real-time data from your machines, physical machines, the actual asset, and you're transferring them to a virtual copy, which is existing in the cloud, which you are calling the digital twin. So once informed with this real-time data, the virtual model or the digital twin can then be used to run simulations, study performance issues, and generate possible improvements, all with the goal of generating valuable insight, which can then be applied back to the original machine.

Mike: That's a great explanation, DP. I love how you introduced the simulation part as well because it really is a simulation model, but with real-time data on something that is actually operating, but exists as a virtual twin of the physical machine, which sounds pretty complicated, but I think you did a really good job of just explaining this is what it is. You've had a lot of experience with digitalization, not only with machines, but with enterprise transformations; do you have any examples of maybe some communications networks or specific sensing capabilities, or even diagnostics, and how those work?

Divya: You mentioned a lot—communication networks, sensing capabilities, diagnostics work—all of those come into the picture. Without going into a more technical side, I'll give you some actual real-world examples where we are, how we are engaging. So we are currently working with a few machine builders who are interested in providing additional services like you mentioned earlier at the very start of this conversation, maintenance as a service.

What we are doing is we are connecting our sensors to an edge gateway that is then transmitting the real-time status to our cloud solution. The machine builders can then monitor the various parameters. And the neat part is these machines are deployed globally, all over the world, but they can still watch these monitors and watch these parameters and the settings. And based on all the different criteria that they have, then they can schedule the field-service team to go out there to fix any abnormalities they're observing or to replenish any stock that needs to be provided because of the levels they're doing.

So what they're really seeing is an additional layer of service that they did not do earlier, I'm doing it now much more efficiently because they have insight. So they're utilizing our products, and they're communicating securely utilizing MQTT protocol. Again, I didn't want to go into too much in the technology side, but there's a lot of these industrial protocols that you're now using for communicating securely between the shop floor and into the cloud. So those are some of the practical experiences we're seeing, and we're doing a lot more of that with the machine builders.

Mike: I can remember 10 or 15 years ago even when machine builders started talking about this availability. Back then, you had to get through a firewall and the connectivity was a bit clunkier than it is today. But even the ability to collect that data and then the potential for a machine builder to sell productivity rather than a capital asset was kind of on the tip of everyone's tongue. Is it happening? Are you seeing more of that starting?

Kaeser Compressors had started doing that years ago where they would sell volume of compressed air rather than the actual compressor. Are you seeing that with other types of machinery now, or is it still kind of working its way into the process?

Divya: Well, the underlying technology is now available, readily available, in fact. And the other criteria was the remote connectivity, as you mentioned earlier. People didn't think it was secure. And with a lot of the secure communication protocols that have come into existence today and some standardized as such with encryption and all the other aspects coming in; plus the biggest concern earlier was that you might send some intellectual property out as you're transmitting data from the machine equipment. And those fears are gone because right now all you're sending is operational data, is the pump running, is the motor running, not running, how long it's been running, if it failed, why did it fail. That really doesn't compromise the intellectual property or product integrity of whatever the customer is manufacturing.

That is allowing a lot more acceptance, and, as you mentioned earlier, the technology has advanced so now it is making it easier, with all the underlying Industry 4.0 technology, the cloud, the edge gateway, and bringing in some computing model all the way down to the edge, doing it as close to the manufacturing process as possible, that delays the latency of actions that is happening on the shop floor. So there's a lot of advancement that has come along that is enabling all this today.

Mike: Fantastic. So, finally, the $64,000 question, doesn't digitalization require machine replacement or even at least an upgrade to what's in operation? How can a machine builder, or a system integrator, or an end user for that matter execute some sort of affordable modification to machinery in existing brownfield locations where there's already an installed base?

Divya: It's a common mistake that we think that the legacy systems that you have must be upgraded or replaced to take advantage of digitalization. You know, we see it all the time. You know, this is a question we get asked. What is typically needed is addition of sensors. Do you have sensors that are collecting the data that you want to see? That becomes the most important aspect. We call it sensorizing the machines. Do you have sensors to pick the values, the variables that you're looking for, the indicators that you're looking for? If you have those today, you collect that data, then the question that comes in is, how do I get the relevant data out of those systems, right?

So, again, you do not have to replace the entire machine. If you have older sensors, if you have older machines, we can still connect to those. We have lots of new gateways that are coming in for transferring from one protocol to older protocols, to newer protocols, and then taking that data into the other, and then transmitting it out using the standard protocols like MQTT and the other ones that I mentioned earlier and taking advantage of the whole cloud solution at the analysis side.

So, if you have sensors, you are already there. The question is now, how do I get data out of the sensors in a format that I can use? So there are technologies, there are products, there are solutions that can do that. But, if you have no sensors, then that's what you need really to sensorize your machines. You don't have to replace the machines or buy new equipment with the new sensors on them, you can sensorize your existing equipment and pull all the data out as needed.

Mike: Great. Excellent point and a great solution. These are surely exciting times. Tthanks, DP. Thanks to all our listeners for joining us on "Control Intelligence," the podcast for "Control Design" Magazine. Thanks, of course, to SICK Sensor Intelligence's Divya Prakash for his insights into machine digitalization. Thanks for joining us, DP.

Divya: Thanks, Mike. My pleasure.

Mike: And if you enjoyed this episode of "Control Intelligence," don't miss our older episodes, and subscribe to find new podcasts in the future. You can find our podcast library at controldesign.com, or you can download all episodes via Apple Podcasts or Google Play. Thanks again, DP. I really enjoyed it.

Divya: Thank you.

For more, tune into Control Intelligence: The Podcast from Control Design Magazine.