As 3D printing grows in popularity, one of its core flaws is beginning to surface increasingly often. The objects these devices create, one layer at a time, tend to fall apart, change their shape, and lose structural integrity over time. Now, experts at Purdue University propose a way of solving this issue.
Scientists say that, most commonly, objects manufactured through 3D printing fail at the points of highest stress. But this problem can be now addressed by using a computer program that automatically makes objects stronger before they are printed.
The software was created by researchers at the Purdue University and Adobe's Advanced Technology Labs. The team was led by Bedrich Benes, an associate professor of computer graphics at Purdue.
3D printers work by constructing objects from scratch, on a layer-by-layer basis. They can use several types of materials, from plastic to metal, but they somehow fail to create items strong enough to withstand normal handling or shipping.
“I have an entire zoo of broken 3D printed objects in my office,” Benes says. “You can go online, create something using a 3-D printer and pay $300, only to find that it isn't strong enough to survive shipping and arrives in more than one piece,” ATL senior research manager Radomir Mech adds.
What the new software does is it first analyzes the structure of the object that needs to be printed. Once this stage is complete, special algorithms figure out which segments are most prone to failure, and selects from a range of possible solutions to this problem.
Most of the times, the software may choose either to increase the thickness of key structural elements, or to add struts, for additional support. The third option is to identify the heaviest elements, and then hollow them out.
“We not only make the objects structurally better, but we also make them much more inexpensive. We have demonstrated a weight and cost savings of 80 percent,” Mech explains.
“The 3-D printing doesn't have to be so precise, so we developed our own structural analysis program that doesn't pay significant attention to really high precision,” Benes adds.
In the future, the team plans to create a version of the software that can be used on 3D printed objects that feature moving parts. These are a lot more difficult to analyze, and the program may have to choose multiple solutions simultaneously to address faults in the original plans.