Modelling and Validation of Laser Powder Bed Fusion for the Optimisation of Laser Parameters

Abstract

Additive manufacturing, and in particular Laser Powder Bed Fusion (LPBF), enables design and manufacturing capabilities that are not accessible through conventional pro-duction routes. LPBF supports near-net-shape fabrication of highly complex geometries and functional integration through geometry-driven design across a wide range of alloys. Despite this potential, adoption in safety-critical and high-value applications remains constrained by limited reproducibility and process reliability. Key challenges include feedstock qualification, control and detection of defects such as porosity and distortion, and achieving consistent microstructural and mechanical properties in final components. Multi-scale physical modelling has the potential to support improved understanding and control of LPBF by providing insight into melt pool behaviour, thermal history, and solidification processes. When integrated with design tools, such modelling may as-sist in identifying processing conditions associated with instability or defect formation. While generative design has advanced rapidly in recent years, the routine integration of physically based, multi-scale modelling into design and optimisation workflows re-mains limited. ESI Group has developed a multi-physics simulation platform capable of modelling LPBF processes from the powder scale to the part scale, including melt pool dynamics and their coupling to thermal and mechanical response. As model fidelity increases, sensitivity to uncertain inputs and modelling assumptions also increases, particularly with respect to thermo-physical material properties at elevated temperatures. A lack of consistent experimental data for validation, especially at the melt pool and powder scales, remains a significant limitation in the literature. Many existing studies focus on single alloys or narrowly defined processing conditions. To address this gap, the present work investigates multiple alloys (Inconel 625 and 718,AlSi7Mg, 316L, 17-4PH, and a high-entropy alloy) processed on a single LPBF system using consistent experimental methods. By combining systematic experimentation with numerical modelling and data-driven analysis, this thesis examines melt pool behaviour and evaluates the capabilities and limitations of modelling approaches used to describe it. Rather than seeking fully predictive capability, the work focuses on understanding variability, identifying robust melt pool metrics, and clarifying how modelling tools can be used alongside experiments to support interpretation and process optimisation in LPBF.