Five years after the emergence of “Industry 4.0” – the concept of connected manufacturing in which networked sensors and assembly systems automatically deliver efficiency and quality improvements – what progress has been made towards the Smart Factory of the future?
What’s in a name?
With more aliases than the FBI’s Most Wanted, it’s no wonder that it’s hard to pin down the reality of Industry 4.0, a.k.a. Smart Factories, a.k.a. The Fourth Industrial Revolution, a.k.a.… But whatever name is used, the common vision is of systems that can sense what’s happening in the world, share that information with their peers, and react in an intelligent way to optimise the industrial process.
The first important component of the concept – process visibility – leverages advanced sensing technologies and improvements in machine-to-machine connectivity to build a “God’s-eye” view of a manufacturing activity. This is Big Data as captured by the Industrial Internet of Things, and lets planners and supervisors understand how their processes are performing – they can see bottlenecks and inefficiencies which were previously hidden and can adjust the workflow accordingly.
But Industry 4.0 goes further, using that data to empower manufacturing automation that can make changes to the process on its own, instantly reacting to production challenges as they arise. For example, a factory which has received a large number of orders for a particular model of product might, within a few seconds, automatically reprioritise resources within the plant to meet that demand. This represents a huge potential increase in the capability and flexibility of manufacturing systems, but also requires those systems to be trusted completely to make the right decisions – the plant’s productivity now depends upon them.
Location, location, location
To rely on automation in this way, the view of the world constructed by systems from sensor data must be complete and accurate. Only then will automated systems have enough information to make high-quality decisions which reflect the choices that well-informed process managers might make in the same circumstances.
Many aspects of the factory environment can be sensed, such as temperature, light level, vibration, and so on. Although these will certainly be relevant to descriptions of specific processes, it’s not always obvious how to build up a coherent picture of what is happening across a factory from these one-dimensional measurements. Instead, some successful Industry 4.0 installations have taken the approach of tracking the precise locations within the factory of relevant items, such as work-in-progress and assembly tools. Different parts of the factory are used for different stages of production, so location immediately reveals the stage of assembly of an item. And when two things are in the same place – a tool and some work-in-progress, say – there’s a good chance that there’s interaction happening between them.
Outdoors, GPS technology works well for accurate location. Indoors, though, it’s a different matter – radio signals are obstructed by and reflect off surfaces like walls and ceilings, making accurate tracking difficult. Fortunately, technologies like ultrawideband (UWB) radio now allow tracking of objects indoors to within a few tens of centimetres. By placing UWB transmitter tags on objects, and mounting networked UWB receivers around the factory, we can locate thousands of objects in a factory in real-time. The most advanced UWB systems can find the 3D position of a tag even when only two sensors detect signals from the tag, by measuring both angle- and time-of-arrival of the signals – this is important in practice because there are so many ways in which the tag’s signals can be blocked in a factory.
The quiet revolution
Commentators often refer to Industry 4.0 as a future technology, but these systems are already in use – for example, around 4% of the world’s cars are now built using Ubisense’s Smart Factory system which gives manufacturers like BMW, Daimler, VW, Honda, Ford and GM the capability of Mass Customisation. (In fact, the first such Ubisense production system was installed in 2008, several years before the term “Industry 4.0” was coined!)
Customers now demand cheaper cars, faster delivery and, crucially, more choice. Henry Ford’s production line of 1913, where every car being produced was identical, has been superseded by one where many different car models, and many variants of each model, are being made on the same assembly line.
Now, when an assembly tool is used on a car, it must be configured to use the exact settings required for that individual vehicle. For example, a line might manufacture cars with a composite bonnet and other cars with a steel bonnet, and applying the torque required to screw in a steel bonnet to a composite part might cause irreparable damage. But by precisely tracking the location of tools and cars in the factory, Ubisense’s system spots when a tool is working on a particular car, and automatically configures the tool for the car it is working on, without operator intervention.
Of course, this is a mission-critical system and production depends on it always making the correct choices. Gaps between cars on the lines are often sub-metre, and tools are used deep inside the cars, making tracking performance paramount. Real-time response is essential, too – an operator can work on the boot of one car at one moment, then turn to work on the bonnet of the next car on the line a second later, and the system must reconfigure itself within that time.
Whole plants now work this way, and scale brings its own challenges. A typical Ubisense installation tracks thousands of cars and tools in a factory, and tags have batteries which last many years, since there are no opportunities to recharge them on a frequent basis. Tags and sensors form part of a large IT infrastructure, and must fit seamlessly into the customer’s existing IT environment, with the requirements for monitoring and upgradeability that that implies. Finally, the software which operates the production systems must be bottleneck-free, supporting these large systems whilst maintaining real-time performance.
Start small, but plan big
With major manufacturers already benefiting from Industry 4.0, what are the barriers to increased adoption?
Firstly, manufacturers are genetically cautious: there are many more ways to break a working process than improve it. Industry 4.0 is a big change – perhaps the biggest since the invention of mass production – which isn’t a bad thing, but it’s a good reason why Industry 4.0 won’t be adopted wholesale, instantly.
Secondly, and perhaps more importantly, assembly plants have historically been driven by the shop floor, with IT providing a supporting function. But Industry 4.0 is the IT-ification of manufacturing, with IT taking a leading role right down to the level of process controls. This is going to involve adjustment and a blurring of lines between departments.
But end-to-end real-time connectedness of all things can be done in stages. Industry 4.0 can be adopted in islands, building local connectivity that doesn’t even need to leave the line, let alone the plant. The best path forward is to avoid the bigger barriers, and quickly implement projects which demonstrate the value of real-time visibility in process control even at modest scale. The next step is to connect the islands once they have paid for themselves with positive return-on-investment. By that point the decision to make a strategic investment in Industry 4.0 won’t be a risk, but an inevitability.
Ubisense’s new AngleID product helps ease the transition to Industry 4.0. A single AngleID sensor can watch for the presence of UWB tags in multiple “spotlight” zones, each covering a specific 2D area in the factory related to an important process activity. AngleID sensors can talk directly with devices like programmable logic controllers (PLCs) and tool controllers, triggering them in response to activity – for example, a wireless tool might be automatically disabled when it leaves the zone in which its use is authorised.
AngleID sensors can be set up in a matter of minutes, and can bring the benefits of location-awareness to bear on production problems, without the difficulties involved with integration into a plant-wide IT environment. But as more islands of coverage are introduced, the same AngleID sensors and UWB tags can be networked together and reconfigured to form a seamless 3D tracking system which covers the entire factory. This “On-ramp to Industry 4.0” is the best of both worlds – immediate benefit with low initial cost, but also future-proofed and scalable.
Frothy or fundamental?
As Industry 4.0 reaches the top of the hype curve, it’s important to understand that some manufacturers have been benefiting from these ideas for several years. Like any new technology, there are challenges to overcome in order to achieve more widespread adoption, particularly in the relatively conservative manufacturing sector, but new products allow step-by-step introduction of the concept, and it seems a good bet that when it comes to factories, Smart will soon become the new Standard.