| Abstract Scope |
In powder-bed additive manufacturing, metal powder defects can compromise product quality. We introduce a non-destructive evaluation method sensitive to thermal properties at adjustable depths. This method modulates the energy source and analyzes temperature responses in the frequency domain, enabling us to assess thermal properties and detect defects in metal powders and printed materials while providing clearer melt pool temperature measurements.
Our experimental setup, revealing distinct responses tied to material features such as core detection, age, oxygen content, and particle size distribution. This real-time detection process works seamlessly with existing powder bed fusion lasers and demonstrates sensitivity across materials like Cu, AlSi10Mg, In718, SS316L, Ti64 G23, and G5. Frequency-domain measurements are less noisy and more robust than traditional thermography methods. By applying machine learning techniques, we successfully identify powder core, age, oxidation, deposition thickness, and size distribution. This innovative approach has promising potential to enhance quality control and process monitoring. |