Development of a self-learning three-roll-bending machine

for adaptive bending of profiles with variable cross sections and contours

The aim of this research project is to develop an adaptive feedback control for the three-roll-bending process and to set up the required hardware and software basics. Especially the spring-back, occurring during bending, which is strongly dependent on the material properties of the profiles, is to be compensated by the desired process control. By assuring quality and productivity, the competitiveness of companies in Germany is stabilized.


Three-roll-bending offers the advantage of a high flexibility of the producible contours and the applicable profile geometries by comparatively low tool costs. However, with the exception of some standard profiles, the bending of radii and contours is controlled manually. Therefore, the profile bending with the three-roll-bending machine requires high manpower and high competence of the operator. A continuous monitoring of the profiles is hard to be realized as well. In particular the variation of the profile characteristics requires a steady adaption of the process in order to assure a constant product quality.

[1] Model of a three-roll-bending process
[2] Process variables


To develop the adaptive feedback control, a numeric model of the process is set up, which is validated by an experimental study of the process. With the help of the conducted simulations, a process analysis is carried out, in order to determine suitable and measurably actuating and control variables. The developed feedback control will be tested in further simulations and after that implemented on a machine prototype and applied to the real process.