Simple method to construct process maps for additive manufacturing using a support vector machine

We propose a simple method to construct a process map for additive manufacturing using a support vector machine. By observing the surface of the built parts and classifying them into two classes (good or bad), this method enables a process map to be constructed in order to predict a process condition that is effective at fabricating a part with low pore density. This proposed method is demonstrated in a biomedical CoCr alloy system. We show that the proposed method is effective at reducing the number of experiments necessary to tailor an optimized process condition. This study also shows that the value of a decision function in a support vector machine has a physical meaning (at least in the proposed method) and is a semi-quantitative guideline for porosity density of parts fabricated by additive manufacturing.