Time-Domain Analysis of Acoustic Emission Signals during the First Layer Manufacturing in FFF Process †

Abstract

Additive manufacturing (AM) has been playing a crucial role in the fourth industrial revolution. Sensor-based monitoring technologies are essential for detecting defects and providing feedback for process control. Acoustic emission (AE) sensors have been used for a long time in a wide range of processes and fields, but they are still a challenge in AM processes. This work presents a study on the AE signals in the time-domain—raw and root mean square (RMS) values—regarding their behavior during the manufacture of a single-layer part in the fused filament fabrication process for two infill patterns. The tests were conducted on a cartesian 3D printer using polylactic acid material. The AE sensor was attached to the printer table through a magnetic coupling, and the signal was collected by an oscilloscope at 1 MHz sampling frequency. It was found that the raw AE signals behaved quite differently not just for the two infill patterns, but within the same pattern. The raw and RMS AE signals contained many spikes along the whole process, but the higher ones were those generally occurring at the end and/or start of a fabrication line. The RMS values, however, were useful for finding the start and end times of each fabricated line for both patterns. The mean RMS values showed nearly constant but distinct averages for the extruder-only, table-only and extruder–table movements.