In 2005, the Lockheed Martin Missiles & Fire Control unit, whose main operations are based in Dallas and Orlando, became intrigued by the aerodynamics of a 50-year-old Swedish-built jet fighter, the Saab A-35 Draken. LM M&FC aerodynamics experts needed a very accurate data map of the aircraft that could be imported into engineering analysis tools to obtain real aerodynamic performance; the reason that they wanted this data was so that they could insure that weapons delivery systems would survive in 21st-century combat environments. Aerospace contractors such as Lockheed Martin continually search for ways to mitigate the high costs of testing. One solution is using a commercial activity instead of a military test facility, and digital rather than physical means.

Complete, accurate surface data of the entire aircraft plus high-resolution scans of weapons and gun bays were required. Lockheed Martin turned to Berding 3D (Santa Fe, NM) for scanning capabilities that would generate surfaces data to be fed into computerized flight simulations. Berding got the job because of its unique experience in long- and short-range scanning, and also because they could react right away and meet Lockheed’s deadline.

Mat Cappel, Technical Manager at Berding 3D Scanning LLC, encountered three scanning-and-digitizing challenges:

• Speed – Lockheed Martin faced 90 days of nonstop data crunching in flight simulations. Initial surfaces data was needed in a week.
• Size of the model – The Draken is over 50 feet long with a 31-foot wingspan and a rudder standing nearly 14 feet high. To minimize the file size, two types of scanners were used: a high-resolution scanner (for highly detailed areas, and a low-resolution (and faster) scanner for flat areas.
• Flexibility – Berding 3D needed a software solution that would process data obtained from both high- and low- resolution scanners.

Solution

In the Draken project, Berding needed to speedily process several gigabytes of point-cloud data in wildly differing resolutions – as much as 10,000x or five orders of magnitude – into a single CAD model. In the high-resolution work, Berding gathered 266 point clouds averaging 250,000 points apiece. This was close-in, short-range work capturing images about two feet square, using a Minolta Vivid 910 scanner.

The low-resolution scans were done with a Cyra2500 from Leica Geosystems. For these scans, the technicians gathered about 20 million points. "This was sufficiently precise for even the smallest aerodynamic shapes but not so high in resolution to capture unneeded data such as rivet heads and hinge points," Cappel said. "The low-resolution work was more like a surveying application for us."

After all the scanning and digitizing – about 250 high- and low-resolution scans aggregating 4.6 gigabytes of data – Berding's final deliverable file to Lockheed Martin was a relatively small 200 megabytes (MB), uncompressed.

Scan Alignment

After the scans were aligned, the resulting point cloud model was transformed in the PolyWorks IMMerge module into a polygonal model in the Stereolithography Tessellation Language (STL) format. PolyWorks creates a polygonal mesh (triangles) adapted to the surface curvature, preserving high resolution across edges and fillets while creating larger triangles in flat areas. Some simulation software packages can process STL files, however, the system used by Lockheed Martin M&FC did not support them. A CAD-usable file was needed.

Creating a CAD-Usable File

Outcome

PolyWorks was also essential to the project as it was able to accurately handle four gigabytes of data. Otherwise, the file would have had to be split into several pieces, requiring additional merging steps and data assembly, doubling and possibly tripling the processing time. PolyWorks saved an estimated two weeks over less efficient software.

Conclusion:

For more information on Berding 3D Scanning visit >www.b3dscan.com For more information on InnovMetric call (418) 688-2061. And for more information on Lockheed-Martin visit >www.lockheedmartin.com