A research team from the NYU Tandon School of Engineering has published a study that uncovers vulnerabilities in the production of carbon fiber reinforced 3D printed parts. The vulnerability is not related to the strength of the parts, but rather in protecting their toolpaths and preventing counterfeit parts.
The ability to 3D print carbon fiber reinforced polymers is creating numerous exciting applications across the aerospace and industrial sectors, among others. The materials are advantageous for many reasons, but their strength-to-weight ratios and durability are most notable. 3D printing, on its side, has enabled far more design freedom for composite materials than was previously possible, and many are now exploiting or exploring the benefits of 3D printing fiber reinforced polymer parts.
However, the process of 3D printing these materials, and specifically the extrusion-based process, can actually reveal the construction of the part and its design. With digital fabrication processes, CAD files and printing toolpaths are valuable trade secrets, so addressing any vulnerabilities is important.
The research coming out of NYU Tandon School of Engineering, which is led by Department of Mechanical and Aerospace Engineering profressor Nikhil Gupta, shows that these toolpaths can be easily reproduced using machine learning tools and CT scanning. This means that part designs could be uncovered, stolen and replicated.
The study, recently published in the journal Composites Science and Technology, demonstrates how these toolpaths can be replicated and 3D printed fiber reinforced parts can be reverse engineered. In the study, the method was demonstrated with a glass fiber composite part, which underwent a CT scan. Machine learning tools were then applied to the microstructures of the scanned model, and the final replicated part reportedly had a dimensional accuracy within one-third of 1% of the original part.
Essentially, the research demonstrates that the direction of the 3D printing toolpaths can be uncovered by analyzing the printed part’s fiber orientation using micro-CT scanning. Machine learning algorithms were programmed to predict the fiber orientation of any fiber-reinforced 3D printed part. These algorithms were validated by the team on cylinder and square-shaped models with less than 0.5° error.
“Machine learning methods are being used in design of complex parts but, as the study shows, they can be a double-edged sword, making reverse engineering also easier,” explained Gupta. “The security concerns should also be a consideration during the design process and unclonable toolpaths should be developed in the future research.”
The study highlights an IP vulnerability that is inherent in the composite 3D printing process with the aim of better preparing part manufacturers in industries like aerospace, which stringently protect their IP and part designs for security and other reasons.