A new artificial intelligence algorithm enables researchers to enhance the efficiency of 3D printing for creating complex structures.
According to a study from Washington State University, published in the journal Advanced Materials Technologies, this advancement could facilitate smoother integration of 3D printing in crafting intricate designs ranging from artificial organs to flexible electronics and wearable biosensors. The AI algorithm was designed to recognize and produce the most optimal versions of kidney and prostate organ models, resulting in 60 progressively refined iterations.
“This allows for optimization of outcomes, leading to improvements in time, cost, and labor,” stated Kaiyan Qiu, a co-corresponding author of the paper and Berry Assistant Professor at the WSU School of Mechanical and Materials Engineering.
The popularity of 3D printing has surged in recent years, enabling engineers to swiftly transform custom digital designs into various products, including wearable technology, batteries, and aerospace components.
However, engineers face challenges in establishing the correct parameters for their printing projects, which can be tedious and inefficient. Factors such as material selection, printer setups, and the nozzle’s dispensing pressure greatly influence the final result.
“The vast array of possible combinations can be daunting, and each experiment demands both time and financial resources,” explained Jana Doppa, co-corresponding author and Huie-Rogers Endowed Chair Associate Professor of Computer Science at WSU.
Qiu has dedicated years to researching the development of realistic 3D-printed human organ models. These models can be instrumental for training surgical professionals or assessing implant technologies, but they must replicate the mechanical and physical characteristics of actual organs, including detailed elements like veins, arteries, and channels.
To optimize the 3D printing parameters, Qiu, Doppa, and their students employed an AI approach known as Bayesian Optimization. After training the algorithm, the researchers effectively balanced three objectives for their organ models: geometric precision, weight, or porosity, as well as the duration of the printing process. Porosity is particularly essential in surgical practice because a model’s mechanical traits can fluctuate based on its density.
“Finding the right equilibrium among these different objectives was challenging, but we successfully achieved optimal printing quality across any type of material or shape,” noted co-first author Eric Chen, a visiting student at WSU who works in Qiu’s group within the School of Mechanical and Materials Engineering.
Co-first author Alaleh Ahmadian, a graduate student at WSU in the School of Electrical Engineering and Computer Science, highlighted that the team managed to evaluate all objectives in a balanced way to achieve favorable outcomes. She also pointed out the advantages of an interdisciplinary approach in their project.
“It’s incredibly fulfilling to engage in interdisciplinary research that leads to real-world applications,” she remarked.
The researchers initially trained their AI program to create a prostate surgical rehearsal model. Due to the algorithm’s adaptability, with minimal adjustments, they could swiftly modify it to produce a kidney model as well.
“This indicates that our method can be employed in the manufacturing of other more advanced biomedical devices and even in various other domains,” Qiu stated.
The initiative received funding from the National Science Foundation, as well as WSU Startup and Cougar Cage Funds.
