PRODUCT REVIEWS
Challenges in the roll-forming process
In the roll-forming process, flat strip material is continuously formed into a defined profile geometry through multiple forming stages. Integrating additional processes such as welding, punching, embossing or forming into the roll-forming line allows a wide variety of profile shapes and sizes to be produced.
“ Roll forming is a manufacturing process that combines several advantages including low energy consumption, low costs per part at high production volumes, high material utilisation and minimal heat transfer into the processed material. The full potential of these advantages can only be achieved under stable process conditions,” commented Dr. Cornelia Tepper and Johannes Hofmann.
“ In practice, maintaining such stability is often challenging. External influences on the process, such as material or temperature fluctuations, can lead to deviations in the profile geometry and consequently to scrap production. The straightening process at the end of the roll-forming line compensates for longitudinal profile deviations such as bow camber and twist. These deviations can arise during processing due to various factors including redundant deformations, changing material properties or incorrect tool adjustment,” they continued.
To this end, the straightening device is adjusted to compensate for inhomogeneous longitudinal strains over the profiles’ cross-section by superimposing a controlled bending deformation. Traditionally, the manual adjustment has been dependent upon the operator’ s experience as they regularly monitored the profile quality. However, due to demographic shifts and other factors( such as staff turnover), this expertise is gradually being lost in many companies. As a result, defects are often detected too late or not at all, which can result in several tons of scrap per day in the continuous roll forming process.
“ Dreistern’ s intelligent straightening device addresses this challenge. Integrated sensors continuously capture reaction forces related to profile deviations. The measured signals are processed in real time by the machine control system and presented on the humanmachine interface( HMI),” outlined Tepper and Hofmann.
“ Should variations in profile quality occur, they are detected by the integrated sensor system and subsequently analysed and visualised through the evaluation algorithms. If the forces exceed the predefined tolerance range, automatic actions are triggered. These range from notifying and assisting the operator to stopping the machine. As a result,
Dreistern’ s award-winning intelligent straightening device.
fewer manual quality checks are required, since the operator is immediately informed when tolerances are exceeded. This approach reduces the risk of undetected quality deviations. The sensor data further allow for prediction of the optimal configuration of the straightening device, since the recorded data responds to adjustments by the device itself,” they added.
Machine-learning assistance
In collaboration with the Institute for Production Engineering and Forming Machines at TU Darmstadt, Dreistern is currently developing a machine-learningbased assistance system to support operators in the precision adjustment of the device.
“ To this end, convolutional neural networks are used to correlate the process forces and the current position of the straightening device with optimal positioning for a straight profile. This technology has already shown impressive results under laboratory conditions. However, the quality of datadriven approaches for industrial use are heavily dependent upon the quantity and diversity of available data sets. To address this challenge, strategies are currently being developed to transfer the knowledge stored in models from different applications to reduce the data requirements for the initial training of models for new use cases in industry. This transfer learning can be implemented for data sets from different profile geometries and materials, as well as for synthetic data from numerical simulations,” confirmed Tepper and Hofmann.
The developed data-driven approaches not only provide the basis for adjustment assistance but also enable the possibility of full automation of the straightening process, in combination with the straightening device.
“ By integrating in-line sensors into the straightening device and leveraging Dreistern’ s evaluation algorithms in collaboration with the Institute for Production Engineering and Forming Machines at TU Darmstadt, the roll-forming process becomes both more transparent and more efficient. The intelligent straightening system enables even less experienced operators to reliably control the complex roll-forming process. Consequently, the applicability of roll forming is further expanded, enabling companies with limited prior experience to successfully implement the process. Rollforming therefore represents a viable solution to contemporary manufacturing challenges and is increasingly recognized as a key technology for future-oriented production concepts and process chains,” concluded Dr. Cornelia Tepper and Johannes Hofmann. n
About Dreistern
Founded in 1949, DREISTERN GmbH & Co. KG is a specialist in roll-forming machinery, units and technology and can provide complete process integration. More than 2,000 units from DREISTERN are in use, producing more than 8,000 different profiles, worldwide.
www. dreistern. com
ISMR December 2025 / January 2026 | ismr. net | 51