Benedikt Waldschmit, M.Sc.

The idea of manufacturing steel components fully automatically has been around for some time. Wire Arc Additive Manufacturing (WAAM) is a process that not only enables the production of customised structures in open installation spaces but also a significantly higher deposition rate than conventional additive manufacturing processes.

As part of the research at the IfSW, an automated process for the inline surface finish of customised WAAM-structures is to be developed by combining additive and subtractive process components.

The research focuses on presenting a design-to-manufacturing workflow for WAAM-manufactured structures, here a steel node to join different standard profiles of closed cross-section or additively manufactured elements. Thereby cross-sectional profiles of different dimensions or geometric shapes (here: square and circle) can be joined together. In addition, the combined process of additively manufactured components and post-processing by removing material to a planned wall thickness using various milling and grinding processes is explained.

One of the most common misconceptions about WAAM is that it is a “plug-and-play” process in which a 3D-model is divided into predefined layers (slicing), which are then converted into motion instructions or G-code. In reality, there are often discrepancies between the design geometry and the actual printed geometry, which can result from differences in layer height or from distortion of the printed structure while being printed.

A digital model contains information about process and input parameters, weld-seam geometries and material properties. The weld paths are generated using partial parametric robot programming (PRP), a coordinate determination method based on mathematical functions. By regularly checking the current height among other things and comparing it with the digital model by the robot controller, the coordinates of the weld path are adaptively updated. The final evaluation of the recorded data, supplemented by 3D-scanning for a target-actual-comparison, is used to generate milling trajectories for a surface finish and to determine, for example, optimized motion sequences or cooling and measuring processes. All stored information results in a rudimentary digital twin.

The node was fabricated by 379 layers (297 vertical; 82 cantilevered). The total manufacturing time was 28.51 h. Of this, only 43.5% were spent on welding directly, 31.9% on tactile measurement, and 20.3% on cooling the structure. The total weight of the node amounts to 20.12 kg. The deposition rate related to the welding time results in 1.65 kg/h.

The recorded data are supplemented by a final 3D-scan model. A point cloud was generated by us-ing a laser scanner and converted into a mesh geometry, which is necessary for performing a target-actual-comparison between the planned and manufactured object as well as generating milling trajectories This was achieved by slicing the mesh model in-to layers with spacing equal to the milling trajectories.

The milling of the specified paths is executed with the cutter aligned orthogonally to the work-piece’s surface and an overlap of the paths of 2 mm to prevent the formation of burrs as much as possible. Nevertheless, small burrs appear between the milling paths, which could be traced back to the non-perfect rigidity, the repetition accu-racy of the robot and the setup’s possible tilting of a few tenths of a degree. To achieve the desired, aes-thetically pleasing surface finish, the milled surfaces were finished with flap sanders.

The project was funded in collaboration with GMSS, ISM+D at TU Darmstadt and spannverbund GmbH in a research project sponsored by ZIM in order to investigate the manufacturing process with regard to technical feasibility and cost-effectiveness.