In-Process Defect Detection for Additively Manufactured Metal Lattices

Direct laser metal sintering additive manufacturing is enabling designers to produce intricate parts that would otherwise be impossible or impractical to manufacture. For example, topologically optimized designs and lattice filled structures are being utilized, and in some cases combined, to reduce overall part weight while maintaining stiffness or other desired performance properties.  However, direct laser metal sintering and other additive manufacturing techniques are generally more susceptible to process variability than many traditional manufacturing techniques. Coupled with the fact that the material properties and geometry are being determined simultaneously in additive manufacturing processes, process variations can lead to both degradation in performance and geometrical defects. 

Despite the advantages that metal lattices can provide to a wide range of industrial applications, there is currently not a clearly defined path towards qualification of these parts in safety critical applications.  Performance of latticed parts can heavily depend on the specific defect population location within the part.  For example, ten broken struts in a large latticed part distributed evenly throughout the part will have a significantly different effect than ten broken struts in a clustered region within the primary load path.  Therefore, qualification of latticed parts will likely ultimately require inspection of every part produced in a lot, not just subsamples from the lot.    

As the size of latticed parts increases, currently inspection technologies will remain limited in their ability to resolve defect deep within the center of the part.  This work proposes an alternative inspection method through steady-state ultrasonic inspection of the part performed during the build process.  An adaptation of acoustic wavenumber spectroscopy – developed at Los Alamos National Laboratory for rapid non-destructive evaluation of structures – was implemented into a direct metal laser sintering machine to inspect production of metal lattices.  The speed of the technique allows for high resolution surface response measurements from steady-state ultrasonic excitation to be collected using scanning laser Doppler vibrometer in between each build layer.  The result of this method is a novel non-destructive evaluation volume for the whole part with consistent resolution within each layer, i.e. maintains high resolution in the center of larger parts.

Focus is given to data analysis and algorithms developed specific to interpreting the steady-state ultrasonic measurements for lattice geometries.  Some examples include: 1) separating out the part response from the unsintered powder within the lattice 2) mapping of response measurements to STL files to locate broken or damaged struts 3) incorporating the geometry into the response measurement interpretation to increase defect detection accuracy.  As lattice structures present very small areas for detection and are surrounded by a large amount of unsintered powder, lattices will be a challenging test subject for this technique. The goal of this work is to show that the high spatial resolution and adaptive nature of the developed algorithms can overcome these challenges and inspire confidence in the direct metal laser sintering production and qualification of metal lattice geometries.