Metal Engineering: turning to artificial intelligence 2 minutes spent reading
Railways

Metal Engineering: turning to artificial intelligence

Christopher Eberl
Christopher Eberl is editorially responsible for the topics on the blog as well as for the apprentice website. With his stories he provides deep insights into the diverse world of the voestalpine Group.

Artificial intelligence is at work almost everywhere in turnouts and rolling lines. Not so long ago, very few people ever thought a turnout could be “intelligent”. But because of digitalization, artificial intelligence, and advances in sensor technology, track control systems have developed an IQ.

voestalpine Railway Systems have played no small role in this development: the intelligent turnout systems produced by the voestalpine Metal Engineering division subsidiary incorporate a wide variety of sensors, providing information about turnout functionality around the clock.

Collecting data on the track

The engineers at voestalpine Railway Systems have set themselves a major challenge for the next generation of turnouts: sensors which divulge the secrets of the turnout interiors and use artificial intelligence to provide more meaningful information. The turnouts would effectively become a local cloud, one which uses edge computing to provide highly accurate forecasting of potential turnout malfunctions and track closures in order to request prompt servicing. This is the current focus of voestalpine research, and experts have been gathering every conceivable type of data on two test turnouts over the past months. They include weather conditions, the quality of the ballast bed, sleeper sinkage, and the constant deformation of the turnout caused by noise. Artificial intelligence is then applied to combine all these factors: it is used to evaluate the data, filter out irregularities, and generate forecasts to ensure the seamless functioning of each turnout system.

The functionality of intelligent turnouts graphically illustrated.

Digital eyes and ears

When it comes to artificial intelligence, voestalpine Wire Technology in Styria is also at the top of its game. The generalist for innovative wire solutions is working simultaneously on several projects designed to optimize its product processes. In one such project, the light pulses captured by a line scan camera create a detailed image of the wire surface. A smart system then uses these images to reliably evaluate whether the wire quality meets the strict specifications—or notifies the team in the control room of the irregularities in real time. Another control unit in the wire rod mill works with acoustic rather than visual aids: artificial intelligence immediately picks up when a change in the typical background noise in a rolling mill indicates a technical problem.

Christopher Eberl