Speaking at an Advanced Manufacturing Research Centre with Boeing seminar, Siemens digital factory chief technology officer Alan Norbury frames the problem that our industry faces: “Technology is advancing at such a rate that it is putting so much pressure on manufacturing that we have to find new technologies and new ways of manufacturing to address this challenge.”

A cluster of technologies that might address some of these issues, and the notion of Industry 4.0, were formulated in Germany in 2011 by Henning Kagermann from consultancy Acatech, together with Siegfried Dais from Bosch. The term itself references previous industrial revolutions – steam power (1.0), mass production (2.0) and computerisation (3.0). The German government, aware of how important export of manufactured goods is to the country’s economy, picked up the theme and rolled it into its High-Tech Strategy 2020 Action Plan. So far, the UK has been more reluctant to jump on the bandwagon, Norbury points out.

Industry 4.0 is a broad church, ranging from process automation to machine management to simulation to even CADCAM integration (which may be something of a stretch) – all are being grouped as Industry 4.0 solutions. At their most basic, many of these concepts share a similar four-step data flow: generate it; collect it; analyse it; act upon it.

In one vision, in the first step, armies of little electronic minions, sensors, are deployed to approximate an objective view of everything happening in a factory at a given time, far more information than would be possible for a worker to see. A case study from IBM solution architect Russ McKay illustrates the idea. There was a problem in a manufacturing company’s paint shop. After cars are painted, extractor fans suck out waste paint. These large fans are vulnerable to being knocked off balance, and during the course of the day they were getting covered with wet paint that sticks to them and then dries. Sooner or later, vibrations in the fan crack the paint, which then falls off onto the top of the car. So the IBM team put an accelerometer on the fan to measure vibration and set up a system to take a meter reading every few hours and record this in an asset management system. As the paint started to build up, they noticed the fan vibration increase too; up to the level that they already knew would be enough to crack the paint and send it flying through the paint shop. Now, instead of sending a worker to repair the damaged paint job (which incurs additional labour costs and expensive production downtime), when the vibration threshold reaches a predefined limit, the system automatically raises a work order for somebody to clean the fan.

In this example, a sensor sent data to a computer that was listening for it, recorded it, analysed it and, ultimately, prompted an action that saved the factory time and money. All without direct human involvement.

This is the promise of Industry 4.0: a sophisticated array of computer hardware and software will keep an eye on things for you and, for example, will warn you when your filters are dirty and need changing, or before that crucial bearing fails and production grinds to a halt.


New super-sensors packed with extra processing power are smart enough to be easier to use than older ones, McKay says: “Now, autonomous measuring systems for physical parameters are standalone digital sensors that are battery powered. You can stick them on the side of a big motor, switch them on and they are able to pick [their signals] up via the internet and monitor them with various types of technology.”

What is clear, even at this early stage, is that this Industry 4.0 is a complicated business, he warns. “There’s quite a lot of setting up on some of this. A lot of people think that you just buy a system and install it and it works. Absolutely not.” For instance, one of the clever uses of Industry 4.0 is tracking, and responding to, the actual condition of the equipment, so that it can be maintained when physically necessary, as opposed to more traditional means of fixing something after it has broken (reactive maintenance) or at some more-or-less arbitrary point (scheduled maintenance).

“For condition-based maintenance, you need to know how a machine performs; to know that you need to know how it has performed; a historical database that you can set your predictions against. And that assumes that you have accurate data: often data collection for machines is not very good; you are missing critical pieces of information for good predictive analysis. Provided you have a good dataset, and that is really important, because most people haven’t, then building a model is not too complicated.”

So virtual models are another part of Industry 4.0. They allow engineers to perfect new manufacturing systems, or enable them to run alternative ‘what if’ simulations of the factory environment, without the massive expense that actually building them in the physical world would require.

Plugging real-world data into a virtual simulation is an especially powerful way to plan factory changes, suggests Norbury at Siemens. “If you have a machine making widget A and you want to produce widget B, which is fairly similar, can you do that?” He answers: “You can test it in the virtual world with all of this physical lifecycle data. That is what cyberphysical systems are about: digital twinning, effectively.”

Siemens’ Plant Data Services offering (08458 507600) relies on comparing physical systems to digital ones. Simulation is used to identify potential problems well in advance to enable corrective measures before unplanned downtimes occur, it says. A related service is Asset Analytics that aims to increase the availability of machines or production lines through continuous online monitoring of sensor data.

IBM (0870 542 6426) calls the computer receiving messages from sensors a ‘message broker’; IBM’s example is called MessageSight. This is the equivalent of the telephone exchange: a machine that receives any number of data streams from individual sensors and can organise them (all of the temperature sensors, or sensors from a particular plant, for example). It can also output some to be recorded in a time-series database. Once data starts to be recorded, it can always be analysed at some indefinite point in the future. “There are an infinite number of things that you could do with the data, once you have got it,” offers McKay.

With lots of pots on the boil, a cook needs to turn down the gas on the one that is boiling over. Two bits of IBM gear do that: in this analogy, the brain is IBM’s asset management software for manufacturing, Maximo, and the eyes are BlueMix, which presents the right data to Maximo from the message broker. Earlier this year, Mitsubishi Electric (01707 276 100) launched its own brains – controller ranges – the Melsec iQ-F and iQ-R, for Industry 4.0-type applications.

Customers may not necessarily need to build a customised Industry 4.0 system from scratch. Below are two systems where much of the set-up has already been done. First, Forcam Force (01606 833 837) is a software package that brings in machine data using the US machine data standard MTConnect, crunches it, provides summary overviews and analyses, and also links in to ERP systems through an SAP software bridge. An add-on is a tool management system from TDM Systems of Germany (0049 7071 9492 0).

Second, this month, Mitsubishi Electric (01707 276 100) launches a new brain, the PMSX micro distributed control system, for smaller-scale applications. This displays a status overview of the entire plant, using graphics, and provides a graphical interface for operators to interrogate details of the system state and make alterations to parameters. It is intended for relatively simple systems that it defines as around 1,000 inputs.

An alternative approach is completely service-based. One permutation of Siemens’ many services for industry takes information from the factory floor via a secure internet connection, performs data processing, and sends business information to the customer.

Siemens is also developing a platform for analysing Industry 4.0 big data, not only for customers but also equipment providers. ‘Cloud for Industry’ will offer original equipment manufacturers – including, possibly, machine tool manufacturers –the capability to monitor the performance of their own equipment, worldwide. The technology is being rolled out to initial customers.

Although these technological developments are only just starting to get traction in UK companies, today’s solutions offer a variety of starting points, ranging from a proverbial toe in the water to diving in at the deep end.


“Industry 4.0 refers to the journey towards self-organising manufacturing operations and a greater distribution of intelligence towards individual machines and components,” says Siemens PLM. “IMI Machine Tools (of Ahmedabad, Gujarat) is an example of one operation to take this approach, reducing product development time and minimising waste, and therefore resource, by upgrading to software that enables an integrated and scalable CAD/CAM approach.”

It designs taps and chucks using 2D CAD on the machine controller, and validated it by machining the part. The process was time-consuming and mostly ruled out complex shapes; results depended on the skill of the designer. The company recently installed Siemens PLM Solid Edge CAD and NX CAM software, and a toolkit for its Mazak Integrex 200ST (01905 755 755). Now, new designs are created in 3D using Solid Edge. Solid Edge files are imported directly into NX, where CNC machine programs are created and verified using toolpath simulation, prior to cutting metal. IMI estimates it saves 500 hours per year in machine programming time since the transition to NX, and it has also seen improved product quality.


German motor and drives manufacturer Wittenstein participates in the German state’s Industrie 4.0 Working Group 1: Smart Factory, along with punching, profiling, bending machine maker and laser specialist Trumpf and automotive OEM Daimler, among others. Its subsidiary, Wittenstein Bastian, is testing cyber-physical production systems in its Fellbach Future Urban Production facility, set up in 2012, in a project funded by the German government. One programme aims to improve the ‘milk run’ of a driver dropping off supplies or picking up finished items within a gear factory by scanning parts into a central database at all stages of production to produce a customised transport route for the driver. Preliminary results suggest that a smart planning algorithm can cut distances travelled by half.

In the UK, aerospace supplier Meggitt began a three-year R&D project last year called M4 (Meggitt Modular Modifiable Manufacturing), backed by the Aerospace Technology Institute, IBM, the UK’s Advanced Manufacturing Research Centre (AMRC) with Boeing and the Manufacturing Technology Centre. Themes include flexible assembly lines, work packages that bring the tools needed for a task that arrive at a workstation, and then instruction at the workbench, including laser and video guides and access to experts housed in a remote call centre. The first phase of the project is the so-called Closed Loop Adaptive Assembly Workbench (CLAAW).

This article was published in the September issue of Machinery magazine.