VDMA honors young talent with “Digitization in Mechanical Engineering” young talent award

The VDMA Software and Digitalization and the VDMA Education Department have awarded the “Digitalization in Mechanical Engineering” Young Talent Award for the 6th time to outstanding graduates from the fields of engineering and computer science. The solution approaches developed as part of the final theses show a high degree of innovation and great practical applicability for the industry.

VDMA honors young talent with “Digitization in Mechanical Engineering” young talent award

2 minutes, 50 secondes

New participation record for nominations
A total of 44 graduates were nominated for this year’s Young Talent Award in the Bachelor’s and Master’s categories by professors from universities in Germany and Austria. “A new participation record,” emphasizes Prof. Claus Oetter, Managing Director of VDMA Software and Digitalization. “Through such close forms of cooperation between industry and the university landscape, we bring new technological approaches and findings more quickly to practical testing and implementation, thus increasing the level of innovation in companies,” emphasizes Oetter.

Dr. Jörg Friedrich, head of the VDMA’s education department, adds: “The variety of submitted theses with a practical focus and the annually growing number of participating universities, illustrate very well that the cooperation between universities and companies is enormously fruitful for the industry.” Around 70 different university locations have already participated in the nomination phases since 2017. “That’s why we as VDMA Education support the Young Talent Award with our Machine House initiative,” says Friedrich.

VDMA honors four young digitalization talents

Jonas Weber, a computer science student at Constance University of Applied Sciences, is awarded 1st prize for the best master’s thesis. Weber wrote the thesis under the supervision of Prof. Dr. Rainer Mueller and in cooperation with Maschinenfabrik Berthold Hermle AG. The thesis aims to show that a complete predictive maintenance solution on a modern machine is possible without the use of external components. For this purpose, historical linear axis measurements were available, which enabled the training of artificial neural networks. The work follows the standard procedure model for data mining (CRISP) in its implementation and delivers as a result a program executable on the machine controller with embedded data preparation, data evaluation and result output.

Simon Griesbeck, student at the University of Applied Sciences Kempten, receives the 2nd prize in the category master thesis. Under the supervision of Prof. Dr.-Ing. Matthias Kuba of the Faculty of Electrical Engineering, he developed a machine learning system at the company AUTEFA Solutions Germany GmbH to optimize the weight accuracy of fiber balers. These high-performance balers are used to compress large quantities of textile fibers so that their transport and storage is more efficient. The developed system allows to intervene in the dosing process in such a way that the deviation from the target weight is significantly reduced. The solution approach can be transferred to many other applications that require high metering accuracy. In particular, companies in the process industry can benefit from the use of a similar machine learning system.

Fabian Kabl, a student at the Ostbayerische Technische Hochschule Regensburg, is awarded 1st prize in the bachelor thesis category. As part of his thesis, which he wrote in the Faculty of Electrical Engineering and Information Technology under Prof. Dr.-Ing. Armin Sehr, he investigated two Long Short Term Memory autoencoders for machine condition monitoring on carousel machines at Krones AG. For this purpose, force data of a clamp star at different rotation speeds were recorded, preprocessed and used to train the neural network. After training, one of the two autoencoders tested was able to correctly classify 100 percent of the data sets in the range of low and medium rotational speeds and 96 percent in the range of high rotational speeds.

Deniz Yesilyurt, mechanical engineering student at RWTH Aachen University, is awarded 2nd prize in the bachelor thesis category. Prof. Dr.-Ing. Thomas Gries and Felix Pohlkemper from the Institute of Textile Technology at RWTH Aachen University supervised the thesis. In cooperation with Heinen Automation GmbH & Co. KG, he developed an AI-based camera system for automated detection and classification of fiber defects in carbon fibers. With a classification accuracy of over 99 percent, the preferred algorithm differentiates between six predefined defect classes and thus enables live evaluation during carbon fiber production. The results of the work are groundbreaking for an efficient and sustainable production of carbon fibers.

More information www.vdma.org