Metal injection molding is a production technology that can manufacture very small parts. Two challenges remain: (1) quality control is cumbersome and (2) energy and material costs are accumulated after injection during the necessary sintering stage. If we detect the quality of a part after injection and before sintering automatically, we will solve both challenges.
A mathematical approach via neural networks was sought to learn automatically the correlation between process parameters that are measured continuously throughout production and the final quality outcome. At the end of the project, this correlation was shown to exist to approx. 96%. Due to this accuracy, good parts can be detected in early stage now. Rejected parts can be removed and do not pass through needless further process steps. This has the effect of lowering the reject rate by a factor of 10. Where previously a production produced about 7-10% scrap, it produces about 1% scrap now. Finally, this innovative method saves material, electrical power and improves the statistical quality of the final shipment.
The German partners in this research and development project was funded by the German Federal Ministry of Education and Research (BMBF) within the Framework Concept ”Research for Tomorrow’s Production“ and managed by the Project Management Agency Karlsruhe (PTKA). The author is responsible for the contents of this publication.