PSI - Issue 77

Roman Hofmann et al. / Procedia Structural Integrity 77 (2026) 237–247 Roman Hofmann et al. / Structural Integrity Procedia 00 (2026) 000–000

238

2

1. Introduction

PBF-LB / M is a versatile manufacturing process that enables the production of highly complex geometries and the tailoring of material properties within a single component. The success of a build, the reliability of the process, and the occurrence of typical defects depend heavily on the chosen build strategies. These strategies determine the local thermal history and influence the microstructure and mechanical performance beyond the adjustment of conventional PBF-LB / M parameters. Build strategies are defined by the applied laser parameters (e.g., power, velocity, and focus), the selected scan strategy (i.e., the spatial motion pattern of the laser), and the coordination of successive layers throughout the build. However, conventional approaches often exploit only a fraction of the available process freedom. Optimization typi cally focuses on parameter tuning or minor variations of standard scanning patterns[9, 4, 12, 16]. This work aims to advance beyond these limitations by developing novel build strategies that exploit the full potential of PBF-LB / M. The overarching vision is to enable more reliable manufacturing with homogeneous material properties, moving closer to the principle of ”First Time Right,” while opening new opportunities for the deliberate design of material behavior and structural performance. Within the scope of this project, the investigated and newly developed scan strategies extend beyond conventional full-hatch concepts. These are based on either the reordering or segmentation of existing hatches, or the new spatial division of the layer. These strategies include:

• Index Reorder: Simple reorder based on the assigned number of the line • Time: Reorder based on the time between new scan path and nearby paths • Pilger: method based on the backstep method from welding. • Voronoi: partitioning strategy - enhancement of the checkerboard scan strategy.

2. Method

2.1. Slicer Framework

In order to implement and evaluate novel build strategies, a dedicated slicing environment was developed, with the majority of the code written in Python and a portion written in Rust. It is important to note that, while Python’s use of the slicer results in comparable slowdowns, this approach nonetheless permits a significant degree of freedom in sub sequent development. The utilization of Rust is contingent upon two factors: the necessity for accelerated processing and the requirement for safeguarding data or code. In the domain of commercial slicers, there exists a notable limita tion on the degree of customizability in terms of modifying scan patterns. In contrast, custom implementation o ff ers an alternative that enables a high degree of control over the process. The framework translates models from .stl or .3mf files into layer-wise laserpaths and allows parametrization down to the level of individual scan vectors. Methods for controlling the parameters within a laser path are also in development. Thereby, machine specific limitations are considered. This high degree of customizability is essential for investigating and developing alternative strategies beyond con ventional parameter tuning. It facilitates the manipulation of each layer, by consideration of geometrical features, and this calculates for each layer if laserpaths should be reordered or parameters changed. Thus, the generation of unconventional scan patterns that cannot be realized with conventional software is enabled. The developed methods described in the following are easy to implement thanks to a python slicer and can also be combined with each other as desired. However, the creation of novel geometric patterns, such as Voronoi, is only possible through the utilisation of a full control slicer.

2.2. Development of Novel Scan Strategies

The full control of preprocessing provided by the custom slicer enables the implementation of diverse scan strate gies that can be derived from di ff erent conceptual foundations, including mathematical optimization, material science driven tailoring, physical process considerations, and programmatic reordering. The main aim here is to achieve ther-