Dimensional Stability of 3D Printed Parts: Effects of Process Parameters

The three-dimensional (3D) printing manufacturing process begins with the creation of a 3D model—using computer aided design (CAD) software—of the part to be printed. Using a type of 3D printing known as fused deposition modeling (FDM®), the 3D printer extrudes molten plastic to scan lines to create individual layers (i.e., the infill): one on top of the other. (Note that "scan" in this context refers to the movement of the extruder head, along an x,y coordinate path, while depositing molten plastic.) This process is repeated until the overall geometry, specified by the 3D model, is built. This process is attractive for producing proof of concept or prototype parts in various fields including automotive, aerospace, and medical. However, FDM subjects the material to rapid heating and cooling; therefore, some degree of undesirable warpage of the part occurs post fabrication. The primary objective of this study was to determine the effect of 4 process parameters (i.e., infill shape, infill density, number of perimeters created per layer, and layer height) on the total dimensional error of a representative 3D-printed part. This part (the "simple part"), used in Trials 1 through 3 of this study, was a square acrylonitrile butadiene styrene (ABS) plate having a nominal measurement of 50 mm × 50 mm × 5 mm thick. A residual error (the difference between the measured post-printing dimension and the theoretical CAD file dimension) was calculated along each given direction and for each test print. Finally, a root mean square (RMS) error (i.e., the square root of the average of the squared residual errors along the length, width, and thickness directions) was calculated for each printed part. Three repeat test prints were carried out for each parameter. The number of perimeters played a key role in the dimensional stability of the part. As the number of perimeters increased up to 5, the RMS error decreased. Beyond 5 perimeters, however, the RMS error increased due to excessive warpage/curvature at the corners of the part. Ultimately, when examined individually, a grid infill shape at 100% density, a 0.4 mm layer height, and 5 perimeters each produced the lowest warpage. In combination, these same 4 parameters also produced the lowest RMS error (based on dimensional analysis of 3 test prints) when used to print a more complicated part (the "stacked part") in Trial 4.