Fourth Industrial Revolution or Industry 4.0 was first mentioned by Bosch at the Hannover Trade Fair in 2011. So, it’s time for an interim conclusion: What has been achieved so far, what does today’s Factory 4.0 look like? What challenges remain? – In this article I will provide a brief description of the conceptual idea “Factory 4.0” (or: Smart Factory) to make sure we’re on the same page. You’ll learn about the awarded Siemens plant in Amberg, and we’ll take a look at the planned Factory 45 by Daimler. Finally, I’ll introduce you to the powerful OpenSource tool StreamPipes, which allows SMEs an easy-to-implement start in setting up their Factory 4.0.
Factory 4.0 – overview
If you’re customer of the German world market leader for wood processing machines, MICHAEL WEINIG AG, your production line throws out ready-to-use end products: Window frames, parquet flooring, furniture. The production process requires wood as an input factor, and the following processing steps are highly automated including dowel-hole drilling, knothole filling, gluing and similar other work steps. The degree of automation will be even higher in a chip factory (because the assembly of printed circuit boards the size of few nanometers requires a precision that is not feasible without machines and high automation).
In short, today’s production already has an impressive degree of automation. Not least, because factories have already undergone several waves of modernization (e.g. “Lean Production”), and yet Factory 4.0 represents a further, significant leap forward in development. How’s that? – Generally speaking, it is all about goals such as flexibility / agility, efficiency (energy, material consumption), or error minimization. And this is achieved by connecting machines, tools, even the products themselves that are involved in the production process and enabling them to exchange data. Every machine, every autonomously moving transport vehicle, every screw to be assembled will be represented by a digital twin in the Digital Factory, and the Digital twin contains all the required information for optimum orchestration of the production process. GPS trackers indicate the location of an autonomously driving vehicle for internal factory logistics, a RFID chip indicates the storage location for a component, the sensors of a machine indicate whether a machine is ready for production or not. The result of such interconnectedness is the Industrial Internet of Things (IIoT).
Factory 4.0 would naturally not add significant value in a single-product company. But a single-product company is a thing of the past, when Fordism and similar concepts prevailed – The manufacturing of today looks different: If you take into account color, equipment, extras, motorization, seat covers and so on, almost every car that rolls off the assembly line today is unique (“batch size 1”). The complexity to be mastered here is enormous. If, for example, a machine breaks down, the question arises: How can material flows, employees and other production capacities be coordinated in such a way that an optimum production result is achieved (taking into account delivery times, internal factory transport routes, etc.). What’s required for such orchestration of the production process is the generation and availability (in real time) of data such as transport times, machine-specific (re)set-up times, availability of service staff, the location of transport machines, etc.. By the way, calculating hundreds (even: thousands) of different production scenarios and finding the (mathematical) optimum is the key strength of quantum technology (or “quantum-like” technology). The Digital Annealer from Fujitsu, for example (described as “The world’s first Quantum-Inspired technology”) was used last year in a pilot project at one of the car manufacturers in Germany, precisely to get closer to such (mathematical) optimum in the production process.
Fig.: Sensor in the production line of a Smart Factory
What technologies are key to realizing the Factory 4.0? – It’s OPC-UA, a standard which enables machines and tools to exchange data with each other. It’s the Blockchain, a technology that allows each product to tell where it was manufactured, when, by whom and on which machines. In order to provide a visionary idea of the future Smart Factory I’ll give the floor to the information scientist Klaus Henning. Mr. Henning provided a truly visionary description of the Factory 4.0 in an article in the magazine FuturZwo:
“The age of machines with their own consciousness begins now. (…) The catalyst is the availability of data thanks to comprehensive networks and digital infrastructures, which are the prerequisite for the effective development of neural networks.
To this end, I would like to give you an example of the ‘intelligent shoe’, which gets an identity even at the time it is ordered. It knows who it is and it also knows who its customer is. He knows what the customer wants from him, whether he should monitor the customer’s parameters, for example. And he also knows what his condition and his path will be: He will have to work his way through production facilities where there is no longer a traditional central control system. The production and transport units are, of course, in turn in symbiosis with their intelligent agents [digital twins] who negotiate with the intelligent shoes. All this could be done ‘democratically’ according to the political principle of separation and interlocking of powers, a method for which there is already a first application for textile weaving machines. (…)
Let us be clear: Products will act as super agents in the future. They will plan their own production and transport. They have requirements for other agents, for example, for a production facility, and they negotiate resources with other agents – on the road or in production.”
Factory 4.0 – The Siemens plant in Amberg
Grey, my friend, is all theory …” (Goethe, Faust I) – Let’s leave the realm of lofty conceptual ideas and get to the factory floor. Let’s take a look at the practice: What has achieved as of today on the evolutionary path towards the age of the Smart Factory (put differently: The Fourth Industrial Revolution)?
As already stated at the beginning, production processes in the electronics industry are already quite highly automated (requirements: precision, purity). No wonder, then, that an electronics industry factory received the Industry 4.0 Award in 2018 in the Smart Factor category, namely the Siemens factory in Amberg. Amberg, this is where the programmable controller for the comprehensive automation solution SIMATIC is manufactured, actually around 16 million devices per year. With this control everything can be automated: From assembly lines to the curtain in a theatre.
Well, why did this factory get the award? What makes this Siemens plant so special? – The answer: A central quality management where data from the entire plant converge in real time. The use of prediction models based on production data which provide early indicators for process deviations; according to Siemens, this can minimize inspection and downtime by up to 40 percent. With the help of more sophisticated AI algorithms in process data analysis, further efficiency potential is to be raised and the causes of faults eliminated before they occur.
It is hardly surprising that data is generated in the order of petabytes: Every single product that leaves the factory (16 million per year) is given its own ID. The data pool provides information about the torque with which a machine has processed a specific component, at which temperature the soldering process took place during the production of the PCB. Actually, the scope of control extends beyond the the value-added chain in the factory to the distribution warehouse: if demand for a specific product increases, production planning is automatically adjusted.
For more information: click here
Factory 4.0 – The Factory 56 from Daimler
The most modern car factory of the world, the Factory 56 of Daimler in Sindelfingen is also guided by the conceptual idea of the Factory 4.0. The inauguration of the assembly hall at the size of 30 soccer fields is planned “at the beginning of this decade” – a precise date is not yet disclosed by Daimler.
One guiding principle is the Industrial Internet of Things (IIoT), which connects selected assembly plants and the conveyor system and enables data exchange between machines and components. For this purpose, a WLAN and 5G mobile phone network is being set up, in close cooperation with the telecommunications company Telefónica and the telecoms equipment manufacturer Ericsson. Approximately 300 autonomous vehicles ensure efficient internal logistics and are integrated into IIoT.
Just as in the Siemens plant in Amberg, predictive maintenance is applied to prevent malfunctions. In addition, data from the supply chain is added to the production control system: Via tracking and tracing , the Factory 56 receives data from part suppliers and can detect deviations along the supply chain – and react more quickly.
It is hardly surprising that Factory 56 means a higher-faster-moving forward. “We will be much more efficient here and need fewer hours per vehicle,” explains Markus Schäfer, the current head of development at Mercedes. What’s amazing, on the other hand: While the factory is also called Fear Factory among blue collar workers, expecting maximum mechanization and minimum staffing, you get to know a stunning fact on the official Factory-56 page: “The know-how, flexibility and high motivation of the employees are the key to success for Mercedes-Benz – especially in times of transformation and digitalization. This is why Mercedes-Benz is reducing the level of automation in the ‘Factory 56’.”
Factory 4.0 – Getting started for medium-sized businesses with StreamPipes
So far, the names Siemens, Daimler – two DAX companies with enormous R&D budgets – have been mentioned. How easily can the vision of Factory 4.0 be implemented by small or medium-sized production companies? In this context I would like to present a powerful open source tool that was developed for AI-based production monitoring and control.
A very well scalable tool, which is mainly aimed at domain experts (Citizen Programmers), who can work on a graphical user interface according to the principle of building blocks. Developed by the Forschungszentrum Informatik (FZI), Karlsruhe.
Fig.: GUI in Stream Pipes
Users set up a workflow via the graphical interface shown here: First, select a data source, such as the real-time data stream from a Siemens S7 PLC (or from a PLC4X adapter, a Bosch XDK sensor, via the MQTT protocol, etc.). Second, assignment of one (or more) data processing algorithm(s). This ranges from simple filters to the use of AI-based algorithms. For example, a pre-trained AI algorithm can evaluate images of a visual sensor to determine whether a manufactured product has a defect or not.
Thirdly, this data can be visualized in a dashboard; the dashboard (as well as the entire application StreamPipes) is made available via a web front-end, so it can be accessed at any time (e.g. by the production manager). Fourthly, Events can also automatically trigger action: This can be the notification of a service technician, or the stop of a machine via signal to the control device. And fifth, you can also persist the data, keyword: data sink, data lake.
There is also a version of StreamPipes that can easily cope with resource constraints, it runs on a laptop for a little “trial and error” (on-premise or on-the-cloud possible). Later, StreamPipes can be run on a NUC or server, and with the help of virtualization (containerization) the application scales very well.