Self-improving AI method boosts 3D printing efficiency

August 22, 2024

(Nanowerk News) An artificial intelligence algorithm could enable scientists to use 3D printing more efficiently to produce complex structures.

A study conducted at Washington State University, published in the journal Advanced material technologies (“Machine Learning Enabled Design and Optimization for 3D-Printing of High-Fidelity Presurgical Organ Models”), could enable more seamless use of 3D printing for complex designs in everything from artificial organs to flexible electronics and wearable biosensors. In the study, the algorithm learned to identify and then print the best versions of kidney and prostate organ models, printing 60 versions that were continually improved.

“You can optimize the results, saving time, money and labor,” said Kaiyan Qiu, co-corresponding author of the paper and Berry assistant professor in WSU’s School of Mechanical and Materials Engineering.

3D printing has become increasingly popular in recent years, allowing industrial engineers to quickly transform personalized designs on a computer into a wide range of products—including wearable devices, batteries and aircraft parts.

But for engineers, trying to figure out the right settings for their print projects is tedious and inefficient. Engineers have to make decisions about things like materials, printer configuration, and nozzle dispensing pressure—all of which affect the final product.

“The huge number of potential combinations is overwhelming, and each attempt costs time and money,” said Jana Doppa, co-author and an assistant professor of computer science at WSU with the Huie-Rogers Endowed Chair.

Qiu has been researching creating complex, realistic 3D-printed models of human organs for several years. They can be used, for example, to train surgeons or evaluate implantable devices, but the models must include the mechanical and physical properties of the actual organ, including veins, arteries, channels, and other detailed structures.

Qiu, Doppa, and their students used an AI technique called Bayesian Optimization to train and find optimized 3D printing settings. After training, the researchers were able to optimize three different goals for their organ models—the precision of the model’s geometry, its weight or degree of porosity, and printing time. The porosity of an organ model is important in surgical practice, for example, because the mechanical properties of the model can change depending on its density.

“It’s difficult to balance all of these goals, but we were able to achieve the optimal balance and get the best possible high-quality print regardless of the print type or material shape,” said first-year paper co-author Eric Chen, a WSU visiting graduate student who works in Qiu’s group in the School of Mechanical and Materials Engineering.

Alaleh Ahmadian, a co-author on the first paper and a WSU graduate student in the Department of Electrical Engineering and Computer Science, added that the researchers were able to look at all the goals in a balanced way, which translated into favorable results, and the project benefited from its interdisciplinary perspective.

“Working on interdisciplinary research by conducting lab experiments to make an impact in the real world is very rewarding,” she added.

The researchers first taught a computer program how to print a model of a prostate surgical sample. Because the algorithm is widely generalizable, they could easily change it with minor tweaks to print a model of the kidney.

“This means that this method can be applied to the production of other, more complex biomedical devices and even in other fields,” Qiu said.