Aligning the holes throughout the 12 layers was tricky. “The first two or three tries, we had some misalignment,” Jadhav says. “But then we got good panels.” Open hole compression testing revealed that the method increased compressive strength 25% to 30% in the adjacent area compared to panels with drilled or water jet cut holes.

In the second phase, the team experimented with a quasi-isotropic lay-up. They used 40 layers of 190 gsm 0/-45 Chomarat non-crimp (NCF) continuous carbon fiber fabric to create the 2.6 mm panels. The team alternated 20 layers of the 0/-45 fabric with 20 layers of the same fabric rotated to 45/90. As before, Jadhav marked the hole locations, carefully separated the fibers and inserted the ¼-inch pin one layer at a time. This time, he also cut the fabric stitches before separating the fibers. It was a time-consuming process. “We had to be careful not to cut the fiber,” he recalls.

Once completed, the panels were tested for compressive strength, static tension and tension-tension fatigue. The results again showed significant improvements in mechanical properties over panels with drilled or water jet cut holes. Microscopic images confirmed that the fibers remained intact.
Jadhav says that future research may focus on creating different hole diameters and automating the process. For now, the team has applied for a patent and is gauging industry interest. Jadhav believes the new technique could benefit many industries, particularly aerospace.

“The future is composites,” he says. “If this technique is used in aerospace whenever there is a joining of two structures, it will reduce delamination problems and help to enhance the life of the products.”

Harnessing the Power of Big Data

Projects: Machine Learning Algorithms

School: University of Washington

Location: Seattle

Principal Investigators: Steve Brunton and Ashis Banerjee

The machine learning industry is expected to hit $9 billion by 2022, up from $1 billion just five years earlier, according to AI Multiple, a technology industry analyst firm. It’s no wonder that schools such as the University of Washington are investing in machine learning research.

The Boeing Advanced Research Center (BARC) in the College of Engineering at the University of Washington hosts joint research projects in which Boeing-employed affiliate instructors work side by side with faculty and students on a range of projects, including ones related to machine learning. Researchers are mining vast quantities of data to find patterns that can be used to accelerate aircraft production rates, streamline design and reduce costs.