Temperature determination in laser materials processing based upon a hyperspectral imaging method of high spectral reconstruction quality

• Applicant: Professor Dr.-Ing. Michael Schmidt
• Subject Area: Joining and Separation Technology
• Term: since 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 326730357

In laser materials processing the knowledge of correct temperature information is substantial. On the one hand, temperature information is of particular interest for the description and characterization of fundamental process domains – such as thermo-fluid dynamics, solidification and evaporation kinetics. On the other hand, temperature information is equally important in application-related areas, for example for defect prevention or quality assurance, process strategy optimization and comparison of experimental and simulation results. However, the acquisition of correct temperature information with high temporal and spatial resolution is yet very challenging. State of the art temperature determination approaches still suffer from inadequate temporal and/or spatial resolution and typically require prior knowledge of the generally only insufficiently known emissivity. Within this project the question will be investigated whether a method for the determination of temperature fields in laser materials processing with high temporal and spatial resolution not relying on a priori knowledge of the emissivity based upon hyperspectral imaging can be realized. Hyperspectral imaging describes an imaging modality where two-dimensional image information as well as the electro-magnetic spectrum for every object-pixel are acquired. Using Planck’s law, temperature information can be derived from the acquired spectra without extensive prior knowledge of the emissivity. Nevertheless, non-scanning hyperspectral imaging requires a reconstruction of the individual spectra, as the spectral information significantly overlaps in the diffraction orders. In this context, there is an underdetermined inverse problem that needs to be well approximated. For that purpose, a suitable hyperspectral imaging system – essentially based upon experiences made in the first project phase – comprising an adapted optical system and a high-speed camera will be realized. The optical system needs to be designed to meet process-specific requirements while integrating as much relevant information as possible for reconstruction of the spectra yet into the image acquisition. Furthermore, a suitable and application-tailored reconstruction approach needs to be implemented. In this context, the existing reconstruction algorithm will be improved to the extent that additional physico-mathematical a priori knowledge can be integrated. That way, an increase in reconstruction quality as well as a reconstruction of full process sequences within adequate time can be expected. A high accuracy in the temperature determination as well as the possibility to deploy the approach in different laser processes requires extensive problem-specific a priori knowledge that is integrated into the algorithm. For that, the a priori knowledge will be derived from experimental investigations of different process scenarios.

• DFG Programme: Research Grants