Engineers using AI removed the errors that occur during 3D printing. New algorithms for machine learning, proposed by a team of researchers at the USC Viterbi School of Engineering, perfectly solve the error problem that inevitably occurs when using 3D printing. This is written on the website of the University of Southern California.
3D printing is often touted as a technology that will be the future for any type of production. Because it allows you to create objects directly from computer drawings, which means that industry can produce individual products on its own, without the use of third-party parts or more personnel. But 3D printing has a high degree of error – for example, often distorted shapes occur in the resulting parts. Each printer is different, and the printed material can be compressed or expanded in the most unexpected way. Manufacturers often have to do multiple iterations of printing before they get the correct object.
Developers from the University of Southern California solved this problem with a new set of machine learning algorithms and the PrintFixer software tool, which allows to increase the accuracy of three-dimensional printing by 50% or more, which makes the process much more economical and stable.
The process that the engineers proposed is called “convolutional simulation of 3-D printing.” The team, led by Qiang Huang, associate professor of industrial and systems engineering, chemical engineering and materials science, received $ 1.4 million in financial support, including a recent $ 350,000 grant. Their goal is to develop an artificial intelligence model that accurately predicts shape deviations for all types of 3D printing and will make it more reasonable.
“So far, we have demonstrated that in printed examples, accuracy can improve by about 50% or more,” says Huan. “In those cases where we produce a three-dimensional object, similar to what was offered to the system during training, the overall increase in accuracy can reach 90%.”
PrintFixer uses data from past 3D printing jobs to train your AI to predict where shape distortion will occur and to correct printing errors before they occur.
The team trained the model to work with equal precision in various applications and materials – from metals for aerospace production to thermoplastics for commercial use. Researchers are also working with a dental clinic in Australia on 3D printing of dental models.