Evaluation of a Cyber-Physical Attack Effectiveness in Metal Additive Manufacturing by Selectively Modifying Build Layer Thickness

To produce functional parts satisfying required functional characteristics, Additive Manufacturing (AM) process maintains a combination of numerous parameters within material-dependent ranges; these include power density, scanning speed, hatch distance, and layer thickness. Unintentional misconfiguration of these parameters is easily detectable as it impacts the entire build. In this paper, however, we consider the case of a deliberate sabotage attack which causes misconfiguration localized to only few strategically selected layers. We propose a method on how such targeted misconfigurations can be executed without hacking into the firmware. Specifically, we altered a build file to mimic localized layer thickness modification by disabling laser beam exposure, while maintaining geometrical and visual part integrity. For two distinct laser powder bed fusion (L-PBF) systems and two metal alloys, we validated empirically the impact of such attack on part quality and demonstrated that it can avoid detection by non-destructive techniques (NDT). The conducted attack illustrates susceptibility of AM to deliberate sabotage attacks and motivates the need of security solutions for this increasingly adopted manufacturing technology.