| Abstract Scope |
INTRODUCTION: Wire-Arc Additive Manufacturing (WAAM) uses traditional arc welding processes, primarily gas metal arc welding (GMAW), in an automated format to produce metal parts with less waste than traditional manufacturing processes. Currently, geometry scanning requires expensive cameras (tens to hundreds of thousands of dollars) and long scanning and data processing procedures (some taking hours to complete). The introduction of scanners like time-of-flight (ToF) cameras have the potential to solve both of these issues.
EXPERIMENTAL PROCEDURE: The printing process begins with a geometry construction in CAD software. This geometry is then imported into slicing software that is then moved into path planning software before being given to a robotic arm that has welding equipment attached. The desired part is printed using as-required feedstock with a clean printing base that is made of a compatible material (such as printing a thin steel 6 inch high wall on a mild steel baseplate). After the manufacturing process is completed, the part is moved from the manufacturing area to an area available for scanning.
After the WAAM process, the part and the build plate are reflective and need to be coated with a dulling spray. The part is then positioned using the software and the image is adjusted using confidence, gain, and exposure features to filter the data capture. The part is then scanned using the Helios2 ToF camera from Lucid Vision Labs. Here, measurements were also made with a structured light GOM ATOS scanner. After measurements, measurement data is used to reconstruct a part for subsequent analysis.
RESULTS: Using the GOM ATOS as the baseline, cost starts at $30,000 and for an accurate model the software requires multiple scans (on a part less than 6 inches took more than 5 scans), each scan taking several minutes to complete. Whereas the Helios2 ToF costs $1500 and a single scan is completed in seconds (could complete a similar number of scans to the GOM in less than a minute).
In both cost and data collection speed the ToF camera has clear advantages but has a disadvantage in data processing and measurement accuracy. While the reconstructions for the GOM scanner are performed in commercially available software, no such software exists for the ToF camera. Thus the raw camera data is manipulated using custom code generated in Python or C++. The degree of difference in measurements is clear when comparing the GOM that has a surface finish of 0.03mm (smooth) to the Helios2 that has surface deviations above 10mm (coarse). The Helios2 ToF camera also has challenges in intensity saturation due to reflectiveness that create data-dropout zones. Adjusting the confidence, gain, and exposure features can mitigate these issues but currently cannot resolve them and requires end-user software development for this camera to be a viable replacement to current metrology scanners in the WAAM field.
CONCLUSION: The use of time-of-flight cameras in WAAM is still early in development with many challenges ahead to solve. Currently, for measurement accuracy, other cameras such as the GOM are much better than the ToF and the quickness and cost of the ToF cannot balance the scale. There is also the issue of intensity blackout of the ToF that needs to be solved either by interpolating or, smoothing of the data. The reconstruction code of the TOF has the potential to solve these issues and must be a point of focus.
KEYWORDS: WAAM, Metrology, Imaging, Time-of-flight (TOF), Scanning |