Automated parameterization of local support at every toolpath point in robot-based incremental sheet forming

Abstract

Within this article, we describe our latest efforts in increasing the geometric accuracy of incrementally formed sheet metal. For this purpose, we design a series of experiments and a special specimen geometry. To enable this series of experiments, we implement a computer aided manufacturing (CAM) system that gives us the necessary freedom in the planning of the tool paths and in the choice of process parameters. With the help of the generated robot programs, we carry out forming experiments whose results are digitized subsequently. From the digitization data thus obtained, we extract the geometric deviation at every tool path point. This deviation is then compiled with various path and process planning parameters to a data set for machine learning purposes. Algorithms that are trained with the help this data set will be able to predict the geometric deviation based the on the given parameters. As a result, it is now possible to automatically compute an optimal parameter combination of support force and support angle that leads to the minimal geometric deviation at every tool path point by what a substantial increase in the geometric accuracy is realizable.