New digital revolutions are constantly changing our lives. While we have integrated smartphones and smartwatches into our everyday lives and the first smart cars are rolling along the roads, industry has also long since entered the next cycle of modernisation. The concept of the smart factory, as the core of an Industry 4.0, is intended to trim the entire value chain to an unprecedented degree of efficiency and flexibility. Will we in future order furniture and cars on the internet from an artificial intelligence and have them made to measure?
The concept of the smart factory is often mentioned in connection with the buzzword "Industry 4.0". This term originates from the German government's high-tech strategy and is therefore mainly used in Germany; internationally it is also referred to as the "Industrial Internet of Things (IIOT)". What is meant is the idea that industry is just reaching a new stage of development: Based on the possibilities of digitalisation, the production process will organise itself to a certain extent in the future.
In this vision, human production managers and machine operators no longer control the manufacturing processes. All technical systems in the factories are digitalised and networked with each other. They have sensors for orientation and can make and implement upcoming decisions in the production process themselves. If logistics, i.e. the delivery and provision of materials and preliminary products, is also organised in a similarly "intelligent" and autonomous way, completely new industrial processes will emerge: At some point, the customer could simply design and order a product on the internet according to his needs - and without the intervention of any human being, an intelligent system takes the order, plans the production, procures the materials, commissions machines with the production and finally delivers a product to the customer that has been made especially for him and according to his specifications. This is the vision - it would probably amount to a new industrial revolution.
At the centre of this revolution is the intelligent factory - the smart factory. The concept is already beyond the stage of visionary fantasy: there are first production sites that are organised as smart factories.
There is no such thing as the "smart factory par excellence". Depending on the industry and product, the level of technical development and the financial possibilities, many different forms of smart factory are conceivable. The question of how much human control one wants to retain over the processes also leads to different approaches. However, the different concepts have some basic principles in common.
If the various systems in the production process - such as machine tools, robots, transport elements and the like - are to work together independently, they must be digitally networked in order to communicate with each other. This is done via ordinary data networks, via WLAN, Bluetooth or also via RFID chips: Workpieces can be fitted with a chip that contains all the information about them - machines read the chip and thus learn what they are supposed to do. QR codes can also be used for this purpose.
But even interconnected robots and machine tools cannot yet act autonomously. For this, they need a kind of digital consciousness - an instance that has all the information about their respective capabilities, their state, their temporal availability and the like - and is able to exchange information with other such instances. Such an instance is called a digital twin. Every object in a smart factory - every machine, every tool and every product to be manufactured - has a digital image of itself in the IT system. These Digital Twins communicate with each other under software control. For example, when a new order is received, the twins of machines can coordinate among themselves which machine is available when for which tasks and which materials are available where. The system plays through different organisational variants and can thus create an optimal plan for production. Only then are the real machines activated and production begins. The digital twin of the product to be manufactured could contain information about what is to be produced, what stage of production the product is in and what needs to be done next.
When machines are not operated by humans, they need sensors to orient themselves. Machine tool sensors, for example, check the condition of the workpiece being machined. Autonomously moving transport elements recognise obstacles and may also be able to orient themselves via GPS if they are to cover greater distances. Containers register their fill level and request replenishment when required production materials run low. Already widespread today are sensors with which machines monitor their own condition. They detect problems and report, for example, when a machine is likely to need maintenance soon. This function is also known as predictive maintenance and can save companies a lot of money: The system always knows the current condition of the machines and can schedule maintenance times in advance so that as little downtime as possible occurs.
In the smart factory, enormous amounts of data are collected - for example, on the condition of all machines and equipment, on the stock and current location of materials, on the products to be manufactured, on energy and water consumption, on orders and the status of their processing. This data makes it possible to optimise the production process down to the last detail and to determine the most efficient method of implementation for each order. However, no human being could do this - which is why artificial intelligence is used in the ideal smart factory. AI analyses data, recognises problems and develops solutions; the system learns to adapt solutions that were developed for old tasks to new problems that arise.
However, centralised control is not always the best way to organise processes. Sometimes it makes more sense to let decisions be made decentrally. The intelligent factory can also do this: if every workpiece in the production process has all the information about itself stored in a chip, for example, then it can communicate directly with the machines without first having to call up information from a central computer: the car on the assembly line tells the robots what they should do next, so to speak.
The advantages of the smart factory are obvious: never before has it been possible to organise highly complex manufacturing processes as efficiently as the smart factory could do. The productivity gain can be even higher if networking is not limited to the individual factory but includes several locations or even several companies. If a factory does not have enough machine capacity or materials for an order at the moment, it is automatically forwarded to other factories in the network or else split up. This concept is also known as shared production.
In general, the Smart Factory makes it easy to further develop and improve production processes. Since all plant data is available in digital form, desired changes can be run through virtually at any time without having to start up a single machine. This digital factory is virtually the digital twin of the real factory and can be used as an experimental field.
But the smart factory is not only efficient, it is also flexible. The interaction of AI and highly flexible controllable machines is also capable of efficiently producing small batches or even individual pieces.
It remains to mention the improved quality control: In a continuously sensor-monitored production, it is easier to detect rejects or even avoid them altogether.
Some things about the Smart Factory still sound futuristic - but there are already locations where the concept is being implemented in one form or another. Daimler's Factory 56 in Sindelfingen is a prominent example. It is a kind of pilot project for the planned transformation of all Mercedes plants.
More often, however, only individual elements of the concept are currently being adopted. Of great practical importance is the area of predictive maintenance. The wind turbine manufacturer Siemens Gamesa, for example, equips its turbines with sensors that monitor the status of the turbines. AI detects potential problems at an early stage and triggers maintenance processes. Elements of the smart factory are also already being used in quality assurance. Bosch, for example, was able to eliminate the problem of undetected production defects with the help of AI.
Every upheaval has its price. First of all, this is to be understood literally. The investment costs for creating the technical infrastructure are high. On the other hand, the smart factory requires significantly less staff in production, which in turn reduces costs.
There is a big question mark behind the issue of security. How vulnerable are huge digital production networks to disruptions, cybercrime or sabotage? Especially when entire corporations network their production facilities, disruptions could have a serious impact on entire industries. Investments in security systems will probably not be the smallest part of the costs in the future.
With security, we are already dealing with a topic that affects not only companies, but society as a whole. This applies all the more to the consequences of Industry 4.0 for the world of work. The smart factory could make many a job superfluous at some point. In that case, we would need new forms of work and employment for millions of people, especially since artificial intelligence could also take over many jobs in other sectors of the economy in the future. Thus, the next stage of industrial development also touches on issues that pose new challenges for us as a society as a whole.
Text: Thorsten Kleinschmidt
Am beliebtesten