The application of additive-manufactured cores and molds is of great interest for complex cast components. Nevertheless, several challenges still exist in utilizing binder jetting in the multi-step additive manufacturing process for foundry applications to its fullest extent. This contribution shows methods that facilitate the use of 3D-printed sand molds and cores in casting series applications. The binder jetting process itself is assessed from an overall process chain perspective to highlight the benefits of its application in series production. The challenges associated with automating mold cleaning for highly complex casting contours are depicted. In particular, employing the method of cleanable mold partitioning is shown to enhance the automation level of the overall process. Mold design tailored to 3D printing is demonstrated to contribute to overall cost and time savings in enhanced core packages. Topology-optimized, lightweight part designs involving complex freeform surfaces may require mold partitioning associated with laborious burr removal processes. A new approach in answer to the shortage of skilled workers in the harsh and hazardous foundry environment is shown. Implementing motion tracking technology is demonstrated to enable economical automated burr removal for minor quantities or high variant diversity in the future foundry. All the methods shown are of great importance for introducing printed core packages into series production.

Complex casting parts rely on sand cores that are both high-strength and can be easily decored after casting. Previous works have shown the need to understand the influences on the decoring behavior of inorganically bound sand cores. This work uses black box and explainable machine learning methods to determine the significant influences on the decoring behavior of inorganically bound sand cores based on experimental data. The methods comprise artificial neural networks (ANN), extreme gradient boosting (XGBoost), and SHapley Additive exPlanations (SHAP). The work formulates five hypotheses, for which the available data were split and preprocessed accordingly. The hypotheses were evaluated by comparing the model scores of the various sub-datasets and the overall model performance. One sand-binder system was chosen as a validation system, which was not included in the training. Robust models were successfully trained to predict the decoring behavior for the given sand-binder systems of the test system but only partially for the validation system. Conclusions on which parameters are the main influences on the model behavior were drawn and compared to phenomenological–heuristical models of previous works.

Quality issues caused by casting defects are commonly complicated to solve, because the part-specific process parameters are not traced to the individual cast part. This problem can be mitigated by the traceability of each cast part with an identifier code. Therefore, a study of the influence of marked surface topography and post-treatments on code symbol quality is desirable for a well-designed traceability system. In this work, the code symbol quality of laser, dot peen, and electrolytic marking methods on three as-cast surfaces of Al–Si alloy, after sandblasting and heat treatment, is evaluated comparatively with a customized image processing software. The result shows that the laser marking method produces the highest performance for different as-cast surfaces; electrolytic marking provides acceptable results only on the smooth surfaces of high-pressure die casting; dot peen marking produces the codes of high symbol contrasts, which are similar to those of laser marking, especially for rough as-cast surfaces of sand casting. However, for all marking methods, the code qualities of all surface topographies decrease substantially after post-treatments. Considering that dot peen marking has satisfying performances as well as low equipment and maintenance costs, this method is more suitable for small- and medium-size foundries to start to trace each cast part in an economical manner.

In the field of casting manufacturing, the post-processing of residual structures and burrs remains a labour-intensive and expertise-driven process, especially for small- and medium-volume production. This paper introduces DemoTeach, a novel semi-automatic (human-in-the-loop) robot teaching method designed to streamline and automate the fettling of complex casting parts using motion capture (MoCap)–based demonstrations. The system enables skilled workers to intuitively guide a dummy tool along casting burrs, generating task-relevant robot trajectories without requiring programming knowledge or extensive 3D data processing. The approach integrates task generation, path planning, and robot control and execution with Cartesian impedance control. A comparative analysis with a reference method based on 3D scanning and point cloud segmentation highlights DemoTeach’s advantages in simplicity, flexibility, and efficiency. Experiments conducted on aluminium casting using a Kuka iiwa 14 R820 (7-axis) demonstrate the system’s ability to reproduce high-fidelity robot trajectories, with end-effector force estimation confirming accurate surface contact during execution. Further opportunities for enhancement are discussed, including quantitative benchmarking, dummy tool adaptation strategies, and the use of augmented reality for real-time feedback and teaching visualization. The results suggest that DemoTeach offers a practical and future-oriented solution for enabling human-guided automation in industrial casting applications.