Advancements in Laser-Powder Bed Fusion: A Study on Machine, Process and In-Situ Monitoring Technology

This dissertation discusses advancements in Laser Powder Bed Fusion (L-PBF) technology. L-PBF is an additive manufacturing technology, where components are fabricated by the addition of material as opposed to material removal as is common in more traditional substractive methods. In L-PBF, metallic components are produced from a metal powder deposited in layers and locally melted by a laser source and a x-y mirror system. As such, components can be produced with a high degree of freedom with little to no impact on the cost. This makes that L-PBF is ideal for highly complex, low series production or highly individualizable components. This has lead to the rapid adoption of L-PBF in demanding sectors such as medicine, aerospace and the oil & gas industry.

In this dissertation a distinctive approach is taken to some of the remaining technological challenges facing L-PBF. In this approach, a machine first perspective is central, with special attention to the design and analysis of the optical train, and the accompanying complex machine/process interactions. By introducing novel designs, but also by clever recombination of disparate technologies and methodologies, advancements are realized in the processing and in-situ monitoring of L-PBF.

Initially, the holistic design approach of the optical train is discussed together with a supporting x-y scanner protocol. The holistic design includes the reverse engineering of closed optical components and quantification of their spectral performance e.g. the f-theta lens. This resulting optical train is the foundation the subsequent results are built upon, with each research subject having a direct link tying it to the optical train. This is followed up with a discussion on the advancements in L-PBF processing. Where the dynamics and processing influence of thermal lensing are discussed and a compensating methodology is presented. Following this, the processing behavior and properties of pure copper produced by L-PBF are discussed. In the process, the mechanism by which highly reflective materials such as pure copper (> 90% reflection) can be processed using conventional fiber laser technology are elucidated. Near dense components are achieved with a relative density of 99.3% and an analytic melt pool model is validated and applied to explain the processing behavior of pure copper. The final part of this dissertation presents some results on in-situ process monitoring. Quality control is one of the biggest challenges facing L-PBF as components are frequently used in demanding applications. This means that stringent safety and certification requirements are typically present and extensive post-process quality control methods are typically required. In-situ process monitoring aims to shift quality control towards an online system, saving time and resources. In light of this, a novel virtual sensing approach is presented, which is able to accurately predict the melt pool depths from coaxial camera images (2.8 % error - optimal range). To achieve this, the melt pool width is extracted from the images and a physics-based analytical model is used to calculate the melt pool depth-to-width ratio from the processing conditions and material properties. Both are combined to predict the melt pool depth. In a second approach to in-situ monitoring, a spatter detection algorithm is applied to the coaxial images and a correlation is found between melt pool spatter and the occurrence of keyhole porosity. These results are validated using both X-CT measurements and high fidelity multi-physics modeling.