Johannes Keppler University in Linz, together with voestalpine Stahl, is launching a series of studies on machine learning and signal processing in the steel industry, Kallanish notes.
The project at the Christian Doppler University Signal Processing and Machine Learning Laboratory at the Institute of the Steel Industry (CD) will last until 2032. The Austrian Federal Ministry of Economy, Energy and Tourism allocated 2.7 million euros ($3.2 million) to him.
Over the next few years, the laboratory will work on developing theoretical principles and algorithms to improve signal processing for monitoring steel production processes.
"Production processes, such as at voestalpine Stahl GmbH, are controlled by sensors, the signals of which are processed by specialized algorithms," says project manager Oliver Lang. He explains that there are different kinds of signals. "One of the most common types of signals is the so-called approximate periodic signals," he cites one example. "Very little research has been done on these signals in the past, and we only had a few algorithms capable of processing them."
Kallanish understands that the reason this happens quite often in strip factories is because of the large number of continuous processes. "Rotations and oscillations create these almost periodic signals, but they also cause a lot of critical interference," Lang concludes.
Author: Christian Kel Germany
Kallanish.com
