An interdisciplinary study from bioengineers at Queensland University of Technology’s (QUT) Center in Transformative Biomimetics in Bioengineering, and Professor Paul Dalton from Oregon University, has outlined the benefits of integrating artificial intelligence (AI) and machine vision into 3D printers for the manufacture of customized medical implants.
Providing its melt electrowriting (MEW) 3D printer with such “eyes and brains” that can essentially see and learn will, the team believes, enable custom-made tissue implants to suit the needs of individual patients. The AI and machine learning integrations will also aid the system’s ability to print heart valves and soft robotics.
“Implants and devices produced through traditional manufacturing methods are usually in standard sizes,” said Center Director and Distinguished Professor Dietmar W. Hutmacher. “However, we have several technical challenges to overcome before 3D printing transforms the advanced manufacturing of implants and medical devices.”
Pairing MEW with AI and machine learning
Through their study, the team is working to bring the benefits of 3D printing to a range of biodegradable medical device designs that have supposedly never been printed before, such as heart valves, bone scaffolds, membranes for dental tissue engineering, and soft robots for minimally invasive surgery.
“Commercially available 3D printers generally offer only high-speed, high-precision, or medical grade biomaterials and rarely do they offer all three,” explained Hutmacher. “This limits their suitability as a manufacturing platform for medical devices such as biodegradable scaffolds for tissue engineering. The addition of AI and machine vision to 3D printing changes this paradigm.”
MEW is a high-resolution additive manufacturing technology that precisely direct-writes small diameter fibers onto a collector. MEW is distinguished from other melt extrusion-based 3D printing technologies by an electrohydrodynamically-stabilized molten jet, which achieves microscale resolutions of the fiber from 0.8 to 50 microns.
The printer’s nozzle is deliberately raised above the collector to provide an ideal visual access point for in-process monitoring during printing when compared to other 3D printing technologies, Hutmacher said.
To provide the MEW printer with the “eyes and brains it has so far been missing” the team is pairing its MEW 3D printer with machine vision and machine learning systems.
The machine vision system comprehensively images the flight path of the extruded fiber to correct errors in real-time, while the machine learning system leverages that information to predict the fiber diameter of the scaffold and make more accurate final products, Hutmacher said.
“The processing of high-performance materials such as medical grade biomaterials for implants is very complex and requires fine tuning of all process parameters which is why we monitor the 3D printing process closely,” he explained. “By using AI, we can evaluate this data stream and identify hidden printing parameter relationships that are not recognizable to humans.”
Leveraging the benefits of machine vision
Looking at metal additive manufacturing’s advancement from prototyping to production, the engineers recognized a key factor in this transition was the implementation of machine vision – an in-process monitoring and analysis method that provides real-time data during the printing process.
For metal 3D printing, this significantly improves the quality and reproducibility of manufactured parts and enables the process to progress through technology readiness levels (TRLs). As such, they hope that introducing machine vision into their MEW system will enable the technique to achieve a level of process control capable of delivering reproducible results and ultimately industrial applications.
During the study, the team demonstrated how machine vision enables the capture and analysis of important and subtle visual information to improve and expand on the precision capabilities of MEW. The MEW jet stability was analyzed using four different electric field environments which led to an improved control protocol of the printer’s electric field for the manufacturing of thicker layer constructs.
The study additionally highlighted the importance of real-time monitoring of the Taylor cone volume, which refers to the cone shape observed in an electrically conductive liquid when exposed to an electric field, in order to better understand, control, and predict printing instabilities for MEW.
The integration of machine vision and AI also moves the technology closer to closed-loop control, accelerating its TRL significantly, the researchers stated.
“This is exactly where the advantage of artificial intelligence lies: it is able to process very large volumes of data quickly, an assignment that is far too monotonous and hence difficult for the human brain,” said Hutmacher. “These new advancements have the potential to transform MEW printers which now will be able to additively manufacture medical devices and implants that have never been printed before.”
Further information on the study can be found in the paper titled: “Convergence of machine vision and Melt Electrowriting”, published in the Advanced Materials journal. The study is co-authored by P. Mieszczanek, T. Robinson, P. Dalton, and D. Hutmacher.
Machine learning in AM
The predictive power of machine learning is being increasingly utilized in many aspects of 3D printing. The technology has previously been deployed to enhance quality control in 3D printing, for the design of new 3D printing alloys, and to better understand the compressive strength of 3D printed construction materials, among other things.
In 2019, 3D printing software developer 3YOURMIND was awarded €1.3 million to integrate machine learning into its additive manufacturing workflow, and 3D printer manufacturer Admatec launched a vision-based process monitoring system add-on for its modular ceramic and metal 3D printer.
More recently, machine learning tools have been leveraged by researchers from New York University’s Tandon School of Engineering to reverse engineer glass and carbon fiber 3D printed components, while a team from Argonne National Laboratory and Texas A&M University have used machine learning to predict defects in LPBF-printed parts.
Looking for a career in additive manufacturing? Visit 3D Printing Jobs for a selection of roles in the industry.
Subscribe to our YouTube channel for the latest 3D printing video shorts, reviews and webinar replays.
Featured image shows AI and machine learning is being integrated into a MEW 3D printer to produce implants for regenerative medicine applications. Photo via QUT.