| dc.description.abstract | Additive manufacturing or rapid prototyping is continuously gaining popularity within and outside the research community. Due to its growing applications and technical and economical advantages over conventional machining, additive manufacturing is being used even to manufacture critical components in aerospace and automobile industry. As a result, there is also an increase in counterfeiting in this technology which can pose a threat to proprietary parts.
This research work deals with authenticating 3D printed parts by comparing the quality of the parts with the precision and bias of 3D printers. The idea is to be able to differentiate between the error distributions of two printers so as to authenticate the parts printed on them. Therefore, the primary research goal of this work is to be able to quantify the differences in error distributions across different printers and characterization of printers. The key challenge is that it is difficult to robustly quantify the difference between two printers simply by comparing their coordinate error distributions. To address this challenge, we introduce a novel topological transformation based on the principle of Voronoi Tessellation, called SplitCode, that exaggerates the differences in coordinate error distributions, thereby, enabling us to better differentiate two given printers quantitatively. Consequently, the method leads to a robust authentication of 3D printed parts.
In this work, through numerical simulations we study the effect of varying mean and standard deviation of error distributions on topological transformation. We find out that maximum exaggeration of the difference between error distributions is achieved on using length, angle and midpoint location to represent the split edge. We present a methodology for quality assessment and authentication and study the effect of different known distributions on authentication. On validating our scheme numerically, we learn that application of SplitCode improves accuracy of authentication between printers with same bias but different precision.
Finally, experimental results show that authentication is possible between two printers with different biases both before and after application of SplitCode. However, authentication of lower quality parts printed on the same printer but at different speeds is influenced by the nature of error distributions of the printer and the print. In certain cases, application of SplitCode is required for authenticating prints that have same bias but different precision. The results of both simulated and experimental studies show that when the quality of the part changes and the problem is to identify a lower quality part, application of SplitCode results in better authentication. This highlights the quality-centric approach of authentication. This research work also offers various opportunities of further exploration in terms of part design, algorithm of SplitCode, imaging and post processing methods and statistical variations. | |
