Cross-sections of CT-scanned 3D Printed Cubes for Porosity Analysis and Prediction

Porosity arises in all additive manufacturing processes. Due to its advantages, FDM is the 3D printing used to generate the coupons, and therefore the datasets for the porosity prediction research. The machine used is a Pulse XE from Matter Hackers. It is an affordable, effective and open-source printer. Table 1 summarizes the nominal ranges of operation of the main 3D printing variables of the XE Pulse.

Table 1 Ranges of Main 3D Printing Properties of XE Pulse

A few variables (see Table 2) have been set as constant given their proven impact or limited one over the print’s quality:

Table 2 Fixed 3D Process Parameters For Coupons Printing

• PLA is a widely used raw material for 3D printing that can also be recycled, and doesn’t require particular storage conditions.
• The recommended bed temperature for PLA is 80C/176F.
• Infill density is fixed at 100% since the objective is to minimize porosity, as it is the main parameter to control the percentage of intentional porosity, mainly for scaffolds and metamaterials where porosity is sought.
• Printing orientation is fixed at the Z-axis as it is the most used print orientation, but changing it and expanding the database could be a topic of future work. 

That leaves the variable process parameters to be the print temperature, the layer height, the infill overlap, the printing speed and the infill type, summed up in table 3

Table 3 Summary of Combinations of Process Parameters for Coupons Printing

The range of allowable values for PLA and used infill types are chosen specifically to not promote active porosity formation.

45 combinations of input variables could be generated. They are then reduced to 11, thus covering the broad ranges of values allowable for PLA, without exhausting every possibility and making it an expensive operation, thus defeating the purpose of machine learning generalization. The final test matrix generated counts 11 combinations, applied for two scales of cubes

22 specimens are generated, as well as 2 more cubes for a repeatability study. 

The main targeted coupon is a 5x5x5mm cube (.2 in) as it is the usual cube used in metal additive manufacturing to gauge porosity and print quality.

A 10x10x10mm cube is also printed in case the capacities of the CT-scanner wouldn’t allow the scan of the smaller one, and for work on the scalability of printing and its impact on the microstructure.