Smart systems, connected and adaptive: the transformation to Industry 4.0 is perhaps the greatest industrial revolution in history. That's because for the first time, it's not just about physical work. Thanks to artificial intelligence, systems are not only digitally connected, but also capable of automated optimization. Dr. Alexander Freytag deals with what this means for manufacturing and the role ZEISS plays in Industry 4.0 every day.
Automation is old hat – even the ancient Greeks tried their hand at it, first and foremost Heron of Alexandria. The engineer is also known as the Gyro Gearloose of the Ancient World. He didn't just invent an automatic temple door. His writings tell us that he also invented the prototype of a steam engine, the enabler of industrialization that changed the world in the 19th century.
Dr. Alexander sits in his office and still deals with automation today, more than two thousand years on. He is an expert in machine learning at ZEISS, a special sub-sector of this technology. And of course, things are quite different to the way they were in ancient times. Artificial intelligence (AI) has changed everything. It is one of the most important and promising technologies of our time. With the help of intelligent solutions, manufacturing processes can be automatically monitored and optimized, adaptively connected and predictively maintained. This increases quality, reduces costs – and thus also prices for consumers. Gone are the days of unnecessary wasted resources, untreated defects and inflexible manufacturing processes as a result. AI thus offers great potential for industry. Freytag and his team are working right in the midst of it all.
Alexander Freytag: a happy coincidence takes him to ZEISS
It was not necessarily something he had expected. At school, Freytag was not interested in artificial intelligence. "I had other things on my mind then," he says, laughing. Nevertheless, he wanted to study computer science; databases and software technology as interfaces to other disciplines appealed to him. Later, Freytag even earned a Ph.D. in AI and actually aspired to having a career in academia. He had no intention of making the switch to industry at that time. Truly. "It was more by chance that I ended up at ZEISS," he says. Freytag has been a machine learning expert at ZEISS in corporate research for more than seven years and says, "It was one of the best decisions of my life." Here, he not only uses his expertise, but also incorporates it into pioneering products and technologies.
Alexander Freytag is driven by the ambition to improve the status quo. ZEISS and the opportunities which arise from artificial intelligence enable him to pursue this goal.
And Industry 4.0 is the playing field on which its solutions are implemented. "We develop technologies that advance our society," he says, also citing a few examples: "We enable processes in production to be optimized in terms of cycle time and quality to be improved. We help medical staff make sound diagnoses. We support researchers to make headway into the unknown. And people have even better vision thanks to our products." As a family man, he is aware of the responsibility involved: "These are the things that make my job at ZEISS unique. Things I can tell my children with a sense of joy and pride."
The smart revolution – Machine learning, automation technology and Industry 4.0
Dr. Alexander Freytag presents insights into how ZEISS uses artificial intelligence
We train computers to have a certain level of curiosity. In other words, they learn in the same way we as parents show our world to our children: this is an elephant, that is a giraffe. And this is a baby giraffe. Similarly, we teach machines the appearance of defects, cracks, and dirt, for example.
Artificial intelligence: Machines learn like children
In order to understand Freytag's ambition, it helps to take a look behind the scenes: in Industry 4.0, everything is networked. Individual steps, which previously were rigidly predetermined, carried out one after the other, and are not interconnected, become processes in which components permanently communicate with each other. This applies, for example, to machines within a production chain or even the entire production. They are tracked by means of sensors. Data is brought together via digital connections and results in a digital image of the systems. At first glance, this is a giant mountain of data. At second glance, it is a treasure that reveals everything about what goes on within manufacturing.
But how do we turn this data into information? As an expert in machine learning, Freytag deals with this question constantly. He says, "When we talk about artificial intelligence in industry, it usually means machine learning." It's about enabling computers to recognize recurring patterns in huge amounts of data – consistently, robustly, and accurately. "We train computers to have a certain level of curiosity," says Freytag: "In essence, they learn in the same way we as parents show our world to our children: this is an elephant here, that is a giraffe. And this is a baby giraffe. Similarly, we teach machines the appearance of defects, cracks, and dirt, for example." While the children learn for life, the machines learn for industry in order to advance it. AI is the technology Freytag uses to enable machines to quickly solve these complex tasks.
Products become more available, higher quality and cheaper
In practise, AI can help to monitor machines, for example, where sensors constantly send information. "However, humans are not capable of constantly controlling 30 sensor values such as pressure, temperature or humidity which a machine records and sends in real time," Freytag explains, "but what we can do is develop algorithms that analyze this data and convert it into information."
In this way, AI algorithms can detect errors in processes long before they lead to a defect and thus to failure of the system. And they can help the machine optimize itself in the next step – in a completely automated way. Furthermore, AI can be used to check not only the machine but also the products for defects. The components are scanned during manufacturing, translated into data and evaluated by algorithms. Faulty parts are removed or repaired. This reduces waste, lowers costs and increases quality. Advantages from which consumers can also benefit: products become more available, higher quality and cheaper especially at a time when life is becoming increasingly expensive and resources are becoming more scarce.
As a high-tech company, ZEISS is harnessing the achievements of artificial intelligence in Industry 4.0. And no two projects are the same: "It's not as if we have a kit of AI solutions that can always be applied to every problem." Freytag and his team develop a bespoke solution for each problem. "The first challenge is often to understand all aspects of the problem," he says: "Only then can we set out to find the right solution."
Industry 4.0 in numbers
U.S. dollars is the figure expected to be achieved by the value creation potential of Industry 4.0 for manufacturers and suppliers in 2025.1
8 out of 10
companies say that Industry 4.0 contributes to sustainable production.2
Finding errors in real time
Here, too, Artificial intelligence, which assists in the processing of almost unimaginably large amounts of data, can be a great help. This is a good fit for ZEISS: "After all, we are active in quite a few areas where large amounts of data are recorded, for example in medical technology, industrial quality control or semiconductor manufacturing," explains Freytag. In the past, he said, ZEISS mainly focused on best-in-class hardware offerings for these markets. Box business, says Freytag. “However, it's difficult to operate such data-producing devices in a way that allows you to draw the right conclusions from that produced data,” the expert said. So, he says, it was natural to say, "We don't just want devices that generate data, but smart boxes that also convert this data into information."
So ZEISS – as an enabler of digitalization – is incorporating AI solutions into its products to accompany customers on their way into the new industrial age. Either tools for software or automation technology are created, or a kind of toolbox from which customers can assemble applications individually. In this way, ZEISS helps manufacturing companies, for example, to detect errors in production processes, monitor machines more efficiently and thus ensure the quality of their components.
For example, a system consisting of sensors, cameras and AI software monitors the additive manufacturing of aircraft turbines or high-tech components for racing. This increases the quality of the manufactured parts and reduces the amount of material used. It also saves manufacturers up to one day of labor time per printing process. AI solutions can ultimately evaluate data faster, more efficiently, and more accurately than a human, and even prompt short-term actions to correct errors. "The solutions are getting better because AI algorithms are learning from the data," Freytag says. Therefore, he expects progress across the board in the future. Especially where a lot of data is gathered – for example in industry. His namesake Heron of Alexandria would certainly have been pleased.
In focus: AI in Industry 4.0
Industry 4.0 refers to intelligent machines and processes in industry, which are connected and adaptive. Things that were previously processed sequentially in pre-defined order now take place in parallel and in an interconnected way. This is made possible with the help of digital information and communication technology. Key technologies for Industry 4.0 are therefore artificial intelligence, digital twins and the Internet of Things (IoT). With global networks across company or country borders, the digitalization of production is taking on a new quality: The Internet of Things, machine-to-machine communication and factories that are becoming increasingly more intelligent are heralding a new era – the fourth industrial revolution, Industry 4.0.
After industrialization, electricity and digitalization, Industry 4.0 heralds the fourth industrial age. Industry 1.0 is characterized by the first automation of processes: machines, driven by steam engines doing the work of humans in factories. It was the start of industrialization. Electricity provided the second step at the beginning of the 20th century. Henry Ford's assembly line was emblematic of Industry 2.0. In the third chapter, Industry 3.0, computers and microprocessors enabled flexible programming of manufacturing machines. And the fourth stage, in which the industry currently finds itself, relates to the development of intelligent, connected and adaptive systems.
The digital digital connection of machines and processes creates countless data sets. For example, machine activities are detected by sensors. Each individual recording is a data set. Translating this data into information – that's what artificial intelligence (AI) is for. That's because these data volumes are far too large for a human to analyze by hand. Artificial intelligence, on the other hand, has the potential to extract information from this data to improve production and service. Possible use cases include predictive maintenance, optimization and automation of processes and quality checks.