PSI - Issue 77

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

240

4

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

next scan closest point next point travel / time previous scan

Fig. 2. Illustration of time-based reordering of a conventional linear hatch.

welds that are performed in the opposite direction of the overall seam progression, thereby reducing distortion and residual stresses while improving seam quality[18]. Transferred to PBF-LB / M, the developed Pilger strategy applies the same principle by segmenting long hatch vec tors into shorter sections, which are then scanned in the reverse direction—from the powder bed toward the previously solidified end as illustrted in Figure 3. This controlled heat input distributes thermal loads more homogeneously, preventing the accumulation of distortion over long continuous scan paths.

segmented line - against main direction 1 2 3 4 5 6 7 8

inital line - main scan direction

Fig. 3. Illustration of Pilger scan strategy. Initial line is segmented into a series of smaller lines, with each segment oriented in a against initial direction.

An additional advantage of this method is its broad applicability: it can be superimposed on any existing linear hatch strategy, since it merely segments the vectors without altering their overall orientation. In the segmentation algorithm, a minimum segment length is defined to ensure process stability while maintaining the intended thermal e ff ect. 2.2.4. Voronoi Voronoi Strategy can be considered an extension of the well-known chessboard scanning concept [13]. This strategy is used to achieve more even heat distribution in order to obtain denser components [13] or to reduce warping and residual stresses[19, 8]. Instead of regular square islands, the cross-section is partitioned into Voronoi cells, which can be adapted in size and position to reflect geometric features or functional requirements. For instance, the core, the edge, sharp corners, thin or bulky geometries, and the downskin region require di ff erent thermal management, necessitating the allocation of distinct parameter sets for each. By varying the Voronoi cell position and their weighting [3], the necessary geometry-specific adjustments can be made. The spatial distribution of Voronoi seeds mainly defines the position and geometry of the Voronoi cells. Seeds can be generated by di ff erent methods like random distribution or ,as implemented in this scanstrategy, by Poisson-disk sampling [14]. In this method, no relaxation [14, 6] step was applied in order to limit computational overhead. Each Voronoi field can be seen as an own polygon and thus has its individual hatching. Beyond defining the hatch ing pattern within each field, the strategy also optimizes the sequence in which the fields are exposed. To avoid random jumps across the layer, a search zone around the last exposed field is defined. The next to exposed field within this zone is then checked according to thermal criteria: if a neighboring field—identified by Delaunay triangulation—was recently exposed, its exposure is delayed to prevent local overheating. If no suitable field is found within the current zone, the algorithm expands outward, initiating exposure in a new region of the layer.