Much of the IT that manufacturers have bought to streamline and improve the design and production process over the past two decades may seem to have reached relative maturity now – computer-aided design and enterprise resource planning, for example.
But things never stand still and the future of “smart manufacturing” has reached an interesting stage as enabling technologies become cheaper. Buzzwords such as “Industry 4.0” or “cyber-physical systems” and “the industrial internet of things” are being bandied about but what do they mean and where are they leading?
To shed light on this, the FT talked to three experts from very different backgrounds: Rahman Jamal, National Instrument’s technical and marketing director for Europe; Matt Davies, director of product marketing at Splunk, the data analytics company; and David Stonehouse (DS), technology and innovation expert at PA Consulting.
FT: What do we mean by the industrial internet of things?
DS: In the internet of things, we have ubiquitous computing and everything is smart and effectively we have a grid of things that are talking to each other. Some big consumer goods companies are looking very hard at this and trying to work out what that actually means at a bottle of shampoo level. You could argue that they’re experts in formulations, they’re effectively “chemical companies” who squeeze some of their knowledge into a bottle and then sell it to you via a retailer. So the link between the manufacturer and their technical knowhow, their competence, and you the consumer is broken and the question is how can the internet of things, or smart packaging, reconnect the original manufacturer with the consumer? This would either be supplying consumer data to help the manufacturer make a better cleaning product, so they’re collecting consumer data, or helping the consumer use that product better because you can actually effectively start to interact with the original manufacturer who has all that knowledge about how to remove the stain from your carpet, for instance.
FT: And what opportunities does this offer manufacturers?
MD: I think it could have quite a profound impact on what we engineer and how we build it, but also on what extra things a manufacturer could offer with their products. Could it be value-added services? If you can monitor a car in real time, can you start to offer predictive maintenance of the car? If you know that the brake pads are thin and they need replacing, or if you know that there’s something wrong with the radiator, if all of these things are connected, then the way that component can get manufactured is different.
Connecting these things more together has an impact on the supply chain, too. Does it mean that the information that we have from the sensors in the things that we’ve made is more real time? And does it give us an opportunity to do something different by having more analysis and insight into what’s going on?
FT: So what is this Industry 4.0 thing? Remind us what 1.0, 2.0 and 3.0 were.
RJ: Industry 4.0 has its roots in computer integrated manufacturing or CIM, but the problem with CIM at the time, in the 1970s and 1980s, was that we didn’t have the internet. So there were proprietary networks, the so-called fieldbuses that connected machines, and that never really took off. Then a couple of years ago the buzzword, Industry 4.0, popped up at the Hanover Fair. It was used by Germany’s chancellor, Angela Merkel, to focus on, secure and underline Germany’s position as one of the leading countries in the area of manufacturing. The aim was to point towards smart factories, which is more of an evolution from my perspective than a revolution, as it has its roots in CIM in the 70s where the third industrial revolution can be positioned (the first being the industrial revolution of the 18th century, based on water and steam power, and the second being the electrically powered production systems of the 20th century).
FT: And where do “cyber-physical systems” fit into this?
RJ: Industry 4.0 is based on cyber-physical systems, a term which sounds abstract but is actually very tangible. Cyber refers to the internet or the interconnection of things or networks, and physical refers to the connection to the real world in the form of sensors, for example, or inputs and outputs, and systems such as those embedded in cars or in fridges and so on. Cyber physical systems terminology can be applied in various different areas of industry or society and smart factory is one application area – others could be smart mobility, smart grid electrical controlled systems or smart health.
FT: One frequently cited example of cyber-physical systems is of a workpiece whose embedded sensors contain the information needed for the machine tool to turn it into a finished part on a one-off basis. How do you see all this developing and what are the implications for flexible manufacturing and mass-customisation?
DS: It’s coming, the cost of sensors is coming down to be able to have manufacturing systems that are flexible because the system is in control. People have spent a lot of time getting manufacturing processes stable and controlled, and effectively testing by batch-testing, or pulling one or two items out of a batch and then assuming the whole batch is correct, or in some electronics there is full inspection, so cyber-physical systems will move into more manufacturing as the cost of testing and sensors, and also the cost of the computing power to actually handle that data, come down. This means that manufacturers will be happier to produce bespoke different items because they can verify that the product matches the order.
RJ: Basically, you gain more transparency in your production process; where a product is in the production line, whether there is a lag in the supply chain, if there are issues in the life cycle management or you need to respond to disruptions and failures in the production side. So all these things can be very well addressed if you base your factory of tomorrow on cyber-physical systems. You end up with more smart products that are more customised, also towards individual needs, and it gives you much more controllability of the process of production.
FT: What are the main obstacles and challenges?
MD: One of the things you’d need to be able to tackle to be able to deliver on some of this promise, is how do you make this secure? If we have all of these things that are connected and we have to manufacture like this and make more of this information available, how do we make sure that information is secure?
The cost, too, is an important factor. How do you do this cost effectively, and not in a 10-year project. What’s the value of doing this at the end, and how do you make it easy enough that everyone can be a part of it, across the supply chain or across manufacturing, but also for the customer as well?
RJ: Managing the complexity of these cyber-physical systems is another hot topic. You have different types of disciplines working together like electrical, mechanical and chemical engineers, you have people looking at the human interface, and so on and so forth, so you require a basis that can be used by all these different user types from an interdisciplinary perspective.
FT: Will there not be huge amounts of data that will be generated by smart factories using these systems?
MD: Depending on who you read, there could be 25bn things connected to the internet in the next couple of years, and even if they only announced a small amount of data every couple of minutes, that’s a huge amount of data. The question that always gets asked around this kind of data is, what’s the value of it? What’s useful and what’s not useful – both now but also historically. If you can keep all of the data around a fleet of a particular manufacturer’s car, if they’ve got 15m cars on the road and you can look, historically, and analyse that data, what trends do you see in terms of faults in particular metal alloys, or whatever it might be. You can start to spot all kinds of interesting trends from that data in real time but also historically as well.
FT: How long do you think it will be before cyber-physical systems, and ultra-flexible smart factories, become pretty much common ground in manufacturing?
MD: I think we’ve seen some already. If you think of the automotive manufacturers, and the high end of manufacturing, we’ve started to see people do that already. They’re pretty sophisticated with what they can do.
DS: I think some of the consumer electronic systems companies are already getting there; you could argue that Dell are quite adept at producing very, very small batches – almost effectively a batch of one – because they’re actually assembling individual subsystems to produce your laptop. Consumer goods are moving into this space, too.
RJ: Some manufacturers of smart devices are already applying these techniques but they are not disclosed because of intellectual property and competitive reasons. Some automotive companies are doing that for their more advanced cars, or some smart device manufacturers are already applying these techniques in their production process.
Industry 4.0 needs more openness right now and more standardisation so that more companies can get involved and so that it does not stay a vendor-specific type of solution.
But it’s already on its way and it’s already being practised, and in 2025 I think we will smile about Industry 4.0 because it will be so common. It’s just like artificial intelligence: every 10 years the definition is a new one but what was forecast 10 years earlier has already been achieved.
For more information on PA technology, click here.