Smart Monitoring
Modern manufacturing processes are characterized by highly dynamic and strongly coupled physical phenomena. Even minor changes in local process states can influence microstructure evolution, defect formation and the resulting material and component properties. Within the Smart Monitoring research area, we develop methods for the systematic acquisition, synchronization and interpretation of process data with the aim of making manufacturing processes quantitatively accessible and assessable in real time.
We use both integrated sensor systems available in industrial equipment and measurement methods developed in-house to capture thermal, optical, acoustic and atmospheric process signals. Our focus is not on data acquisition alone, but on the physically grounded reconstruction of actual process and material states from complex and often indirect measurement signals.
A particular focus lies on additive manufacturing processes, especially Laser Powder Bed Fusion of Metals (PBF-LB/M). Here, we investigate the relationships between process signatures, transient process states and the resulting component quality. Our goal is to identify instabilities, critical process conditions and defect formation at an early stage and to establish robust relationships between process signals, microstructure evolution and component performance.
To achieve this, we combine experimental investigations with time- and frequency-domain signal analysis, data-driven evaluation methods, model-based state reconstruction and artificial intelligence. By integrating multiple sensor sources, we create digital representations of the process that provide substantially deeper insight into manufacturing dynamics than the isolated analysis of individual measurement signals.
The resulting insights provide the basis for data-driven quality assurance, adaptive process strategies and future closed-loop control concepts. In the long term, our objective is to establish a fully sensor-enabled and quantitatively describable process chain in which relevant process states are continuously interpreted and made directly available for process control.
