Researchers at Washington State University (WSU) have used an AI technique to train and find optimised 3D-printing settings for highly complex designs. Their self-improving AI method could improve 3D printing manufacturing of intricate structures for flexible electronics, wearable biosensors and other complex patterns.

As part of the WSU study, the algorithm learned to identify, and then print, the best versions of kidney and prostate organ models, printing out 60 continually improving versions.

“You can optimise the results, saving time, cost and labour,” said Kaiyan Qiu, co-corresponding author on the paper and Berry Assistant Professor in the WSU School of Mechanical and Materials Engineering.

For engineers, establishing the correct settings for 3D printing projects can be cumbersome and inefficient. Engineers have to decide on materials, the printer configuration and the dispensing pressure of the nozzle. All of which affect the final product.

“The sheer number of potential combinations is overwhelming, and each trial costs time and money,” said Jana Doppa, co-corresponding author and Huie-Rogers Endowed Chair Associate Professor of Computer Science at WSU.

Qiu has done research for several years in developing complex, lifelike 3D-printed models of human organs. They can be used, for instance, in training surgeons or evaluating implant devices, but the models have to include the mechanical and physical properties of the real-life organ, including veins, arteries, channels and other detailed structures.

Qiu, Doppa, and their students used an AI technique called Bayesian Optimisation to train and find the optimised 3D-printing settings.  Once it was trained, the researchers were able to optimise three different objectives for their organ models—the geometry precision of the model, its weight or how porous it is and the printing time. Porosity of the organ model is important for surgery practice, for instance, because the model’s mechanical properties can change depending on its density.

“It’s hard to balance all the objectives, but we were able to strike a favourable balance and achieve the best possible printing of a quality object, regardless of the printing type or material shape,” said co-first author Eric Chen, a WSU visiting student working in Qiu’s group in the School of Mechanical and Materials Engineering. 

Alaleh Ahmadian, co-first author and WSU graduate student in the School of Electrical Engineering and Computer Science, added that the researchers were able to look at all the objectives in a balanced manner for favourable results and that the project benefited from its interdisciplinary perspective.

“It is very rewarding to work on interdisciplinary research by performing physical lab experiments to create real world impact,” she said.

The researchers first trained the computer program to print out a surgical rehearsal model of a prostate. Because the algorithm is broadly generalisable, they could easily change it with small tunings to print out a kidney model.