Detecting anomalies in melt-extruded 3D printed parts using in situ data

The methodology involved an introduction of a material defect by varying the filament diameter of PETG filament and measuring the filament diameter values for the length of filament needed (9m) for a full cylinder three times before and after the defect was introduced using two orthogonally-positioned hall-effect sensors and a digital rotary encoder used to track the position along the filament of the cross-sectional diameter measurements. The data collected from the aforementioned process is not included in this repository, but the process preceded the creation of a defect that was both detectable in the in situ data and resulted in a printed cylinder with a statistically significant increase in its pressure decay rate. Following this, six cylinders were printed on the same build plate and real-time in situ data from various sensors equipped on the desktop printer were gathered with the objective to map sensor data onto 3D coordinates to identify on which parts the identified anomalies landed. Determining the effect of filament diameter on the print quality involved pressurizing the printed cylinders with the introduced defect to 1 PSI, then evaluating for statistically signifcant increases in pressure decay rate of the cylinders. As multiple parts were printed together layer-by-layer (opposed to part-after-part), the size of the anomaly was limited to be certain that no more than three parts out of six would contain anomalies.