Digitization in the here and now with Dr. Sophie Wei
Digitalisation and industry 4.0 | People at thyssenkrupp | Intelligent technologies are increasingly smiling at us from futuristic images as a vision of tomorrow. Yet the current use of data analytics and artificial intelligence is already in full swing today and is anything but old news.
We sat down with Dr. Sophie Wei, Head of Analytics & AI, Managing Director & CTO CarValoo GmbH from thyssenkrupp TechCenter Control Technology, to talk about the different application areas of data analytics, artificial intelligence and machine learning.
How data analytics and AI can make industry better
The application of smart technologies in industry is diverse, Wei explains, “With the ability to collect data, plants and machines are generating much more data today than in the past. We collect this data, visualize it, and create statistical trends to detect patterns and identify trends.”
Only after this data collection do the actual machine learning and artificial intelligence come into play: Based on the identified patterns, Dr. Wei and her team of eight build an AI model that can predict trends or issue recommendations. In this way, sources of error can be avoided within individual machines, but also interrelated process chains, and optimization can be achieved.
However, technical progress by no means replaces experts, Wei explains. Rather, data analytics supports and facilitates their work. “In the industrial and plant world in particular, a lot used to be adapted and changed by expert knowledge. Our developments are more or less a coming together of this human expertise and the data,” says Wei.
What is the importance of digital expertise for thyssenkrupp?
“The biggest advantage of data analysis is that we can make the products and applications that thyssenkrupp already offers better,” explains the expert. That’s because with the help of data analytics, the experts can evaluate large volumes of data many times faster than humans ever could. The digital solutions support employees in the plant by making information about the production process tangible that was previously hidden.
DISCLAIMER: These pictures were taken before the outbreak of COVID-19
“Following the data analysis, the expertise of the plant employees is in demand,” says Wei. What correlations and conclusions can be derived from information? Once sources of error and optimization potential have been identified, machines and entire machine parks can be reprogrammed with the help of machine learning. Errors are corrected and improvements integrated.
In this way, not only does the quality of the products continuously improve, the plant also produces less scrap. That makes a lot of difference, because less scrap equals savings in time, costs and resources. But avoiding machine breakdowns and being able to better plan maintenance work is also a result of digital expertise in the industry. Delivery bottlenecks can thus be largely avoided.
Fascination with AI and its future in industrial production
What fascinates Dr. Sophie Wei about working in the field of data analytics and AI is the broad scope of applications. “What I particularly enjoy is the wide spectrum in which we work,” she says. “Not only in the sense that we use different methods and learn something new every day, but also because as part of the thyssenkrupp family, we can create developments for a very diverse and wide-ranging industrial sector.” At thyssenkrupp, her team is involved with CarValoo in developments for the automotive industry as well as in plant engineering. Specifically, that means in chemical plant engineering and mining and working with open-pit mining machines or cement mills.
Looking to the future, Dr. Sophie Wei is certain: Artificial intelligence will not only be commonplace in industry in the next ten years, but will also find its way more and more into products in our daily lives: “I believe that data and aspects of AI will be in every product in the future, whether from thyssenkrupp or in products that you use every day as a private individual. This means that all processes will at least be data-based, if not completely.”
The expert sees one advantage in particular in this development: Products and services will be much better tailored to users. However, she rules out a noticeable change to the smart future: “We will not perceive artificial intelligence as a new technology, but rather as an improvement of an existing technology. The self-driving car – something created entirely from AI and data – is still many years away. Until then, we’ll always have products that have data and models running in the background, but that just feel to us like an improved version of what we already know.”
For further reading: Another interesting project of our digital experts around Dr. Sophie Wei is CarValoo the X-ray glasses for cars.