Experimental evolutionary optimization of geometric integrity in Fused Filament Fabrication (FFF) Additive Manufacturing (AM) process

The objective of this thesis is to optimize the Fused Filament Fabrication (FFF) Additive Manufacturing (AM) process to produce geometrically accurate parts based on two process parameters, namely, extruder temperature and infill percentage with ABS polymer material. This is a practically significant problem, because, poor part dimensional integrity is a critical impediment to application of AM processes in industry. The figures shown below display warping in the part produced by varying infill percentage. It can be visually noticed that as infill percentage increases from left to right, the warping increases too. Two technical challenges in this context are: (1) finding the optimal process window; and (2) assessing the geometric integrity with quick, non-contact approaches such as laser scanning, as opposed to tedious co-ordinate measurement. In order to address these challenges, a test artifact based on the NAS 979 standard is suggested and an evolutionary design of experiments schema is executed in six phases including, central composite and full-factor testing plans. The parts (over 130 test parts) were assessed for geometric integrity in terms of datum independent Geometric Dimensioning and Tolerancing (GD&T) features as per ASME Y14.5-2009 using a laser scanner. The point cloud data (> 2,000,000 3D data points per part) obtained from the laser scanner are used to ascertain geometry integrity (GD&T) parameters including, flatness, circularity, cylindricity, thickness, root mean square deviation (from CAD data), and in-tolerance percentage. Response surface statistical analysis is used to determine the significance (α = 10%) of the effects of infill percentage (in the range of 70% to 100%) and extruder temperature (in the range of 220 °C to 240 °C) on the aforementioned GD&T parameters. Based on the evidence from this data, infill was found to have a statistically significant effect on geometric integrity, whereas, extruder temperature was largely insignificant. The analysis points to an optimal process window of 70 to 80% infill percentage.