Adaptive Bending of Aluminium Extrusions Using an Automated Closed-Loop Feedback Approach

Abstract

A new rotary, compression bending set-up with automated closed-loop feedback control is being developed. The overall goal is to improve the dimensional accuracy of formed shapes using elastic springback compensation. In-process measurement data are transferred into an algorithm (steering model) for prediction of springback and bend angle prior to unloading. Emphasis was placed on developing a physically-based steering model. More than 150 bending tests of AA6060 extrusions were conducted to demonstrate the capability of the technology. Prior to forming, the material was exposed to different heat treatments to provoke a range of stress-strain behaviours, which are known to affect elastic springback. An optical measurement procedure was established to determine key dimensions and their associated statistical distributions. When using traditional compression bending, the results show that the variability in springback of a selected reference point was in the range of 10% of the nominal springback. Using the closed-loop feedback system, the corresponding variability in springback was in the range of 3%, representing a factor-three improvement in terms of dimensional process capability (C p ). It is concluded that the present technology has a high industrial potential, in particular for volume components with tight dimensional requirements.