In-situ monitoring of laser powder bed fusion applied to defect detection

Doctoral thesis, 2023

Additive manufacturing technologies, particularly laser powder bed fusion (LPBF), have received much attention recently due to their numerous advantages over conventional manufacturing methods. However, the use of LPBF is still quite restricted, mainly due to two factors: its typically low productivity, which makes the technology less competitive in applications with moderate to high production volumes, and its limited reliability, particularly relevant for applications where high performance is required from the materials.

The issue of low productivity is addressed in this thesis by adjusting the main LPBF process parameters. An equation for the build rate was formulated based on these parameters, determining their contributions and enabling strategies for build rate maximization. The changes in microstructure and defect populations associated with increasing productivity were determined.

The reliability issue was explored by investigating defect formation, detectability and mitigation, since a major factor compromising reliability and materials’ performance is the presence of defects. Internal defects were deliberately created in LPBF-manufactured material to assess their detectability via in-situ monitoring. Two main routes of deliberate defect formation have been identified while preserving defect formation mechanisms; therefore, this thesis can be divided into two parts according to the approach employed to create defects.

Defects are generated systematically if suboptimal process parameters are employed. The types, quantities, and sizes of defects in nickel-based alloy Hastelloy X resulting from varying processing conditions were thoroughly characterized. Analyzing data obtained from in-situ monitoring made it possible to distinguish virtually defect-free material from defective material.

Defects are generated stochastically due to the redeposition of process by-products on the powder bed. With the aid of in-situ monitoring data, the presence of these defects can be inferred from the detection of the process by-products responsible for their formation. The comparison of data obtained in-situ with data obtained through ex-situ material characterization allowed determining how precisely detections corresponded to actual defects. The impact of these defects on the mechanical properties of Hastelloy X was assessed. A couple of in-process mitigation strategies were investigated, and their performances were evaluated.

By establishing means to use LPBF process monitoring to distinguish high-quality from defective material and detect random, unavoidable defects, this thesis enables the prediction of LPBF material quality. It creates conditions necessary for the first-time-right production of defect-free material at increased build rates.