AM Heat Exchangers for Aerospace Applications
Heat exchangers are critical parts for the aerospace industry. The traditionally manufactured part uses plate-fin brazed and welded assemblies, which result in known thicknesses and roughness for quality departments. When Raytheon Technologies explored using additive manufacturing to replace existing fan duct heat exchangers, the multinational aerospace and defense manufacturer encountered a few challenges.
While AM can create monolithic parts that simplify production and reduce volume, the varied thickness and roughness of the thin walls can degrade the business case for AM-produced parts. And the surfaces in question are not visible for inspection, and roughness changes depending on build orientation. To move to non-traditional designs enabled by AM, Raytheon first needed to prove the process can produce smooth internal walls of a specified thickness.
How could Raytheon achieve the data completeness and quality needed for AM parts? Zeiss Group provided the solution with industrial computed tomography and X-ray microscopy imaging to non-destructively inspect an AlSi10Mg heat exchanger’s internal features.
AlSi10Mg is an aluminum alloy that combines silicon and magnesium as alloying elements. The result of the combination is lightweight parts that are significantly stronger and harder than other aluminum alloys. This toughness makes AlSi10Mg ideal for aerospace parts manufacturing.
Raytheon used a Zeiss METROTOM X-ray CT system to perform an overview scan of the heat exchanger at 60 µm per voxel to check wall thickness and gaps or cracks in the thin walls. To further enhance ROI, higher resolution scans were completed at the top and bottom of the part at 15 µm per voxel and 3 µm per voxel on a Zeiss Xradia Versa 620. The higher-resolution scans verified the wall thickness and allowed the surface roughness to be measured.
Once non-destructive inspection scans were completed, the 3D-printed fan duct heat exchanger was sectioned to reveal the same surfaces for conformational confocal microscopy of matching top surface (up skin) and overhanging surface (down skin). This excavated section inspection acts as a ground truth data for this study.
Once the X-ray CT scan data is aligned, matching surfaces were extracted for surface roughness analysis. The results were compared to a similar extracted region imaged with confocal microscopy. On the smoother up-skin surface, the arithmetical mean (Sa) and lowest valley (Sv) values aligned well. But to pick up the highest peaks (Sp), a voxel size of 15 µm or smaller is required to match the confocal microscopes analysis. This discrepancy occurs because the surface is made up of smooth overlapping meltpool tracks left on the surface by the final pass and with the occasional sintered powder particle to make up the highest peaks.
When the voxel size is greater than the expected powder size distribution, the analysis cannot be trusted to pick up the powder sized peaks. On the down-skin surfaces, the Sa and Sv values increase as expected for AM overhang surfaces. These surfaces contain large powder conglomerates and single powder particles, which leads to highly irregular surfaces with large peaks and valleys.
AM can be used to produce fan duct heat exchangers if build orientation is considered. Parameters must also be optimized for the overhanging thin wall, which cannot be supported or accessed for polishing. CT combined with XRM is a powerful tool for getting feedback during parameter development and for inspection of final parts.
Connect With Us
