Steel Industry 4.0
Efficient material and process development through simulation
The mechanical properties of steel are determined to a large extent by its microstructure – which, in turn, depends on the rolling process and the individual rolling parameters. The targeted combination of these parameters makes it possible to improve the quality of a steel product. An efficient and exact heat treatment, for example, also shortens the process chain. In the first step it is necessary to identify influencing factors. This is followed by an analysis of what effects these have on the microstructure. However, the roller tests required for this process generate an enormous organisational and technical workload.
This is why Georgsmarienhütte GmbH has spent the past few years working with external project partners to develop a fast simulation program. Using semi-empirical models, this program can predict the mechanical properties of the end products based on their chemical composition and the rolling parameters.
How does this work? A special concept was developed in order to exploit the full potential of the simulation software. This takes the form of a fully integrated system for predicting the microstructure and material properties of hot-rolled steel bars in real time.
Using fast algorithms and detailed data storage, the microstructure properties for a flexible number of segments per ingot are calculated based on the real process parameters. The simulated and the measured data are then automatically adjusted and calculated once again.
Using data mining to advance towards Steel Industry 4.0
The accuracy of the prediction is essentially determined by the quality of the process data. The semi-empirical models used to calculate the product’s microstructure are based on the laws of physics. However, they cannot take probability distribution into account. As such, the traditional analytical method alone is not sufficient to fulfil Georgsmarienhütte GmbH’s requirements in terms of precision. Our scientists and developers had to find a new method and link it to the process.
The solution: various data mining processes were integrated into the overall system (data mining is the systematic application of statistical methods to large databases). These processes can counteract and more accurately predict the stochastic fluctuations in the process parameters. The long-term aim is to develop this into a hybrid, self-learning overall system.
New levels of human/machine interaction
The processes of this Steel Industry 4.0 project are very complex. But humans are - and always will be - the link between the analogue and the digital worlds. The algorithm developers are constantly exchanging data with the computer-based data mining models. This enables new methods to be integrated into the concept online, which leads to a continuous improvement of the sensor landscape at Georgsmarienhütte GmbH.
The new expansions to the data infrastructure have laid the foundations on which we can build this digital concept and many more like it, thus allowing us to implement our vision of the future: Steel Industry 4.0.
The quality and the material properties of steel result from a number of reciprocal effects. Georgsmarienhütte GmbH is developing sustainable and efficient materials using a hybrid simulation method.