The powder–gas jet stream (PGJS) is one of the main factors that governs deposition quality in laser-powder directed energy deposition (DED-LBp), yet no standardised method exists for geometrical and dimensional characterisation of the PGJS features. This paper presents a statistical analysis of PGJS geometric descriptors extracted from dual-view optical systems, a lateral off-axis and a coaxial imaging system, under a centred central composite design of experiments varying shielding gas (SG), carrier gas (CG), and powder feed rate (PFR), using SS316L powder.

Comparison of descriptor means and variability across parameter levels reveals a consistent parameter hierarchy: SG is the dominant driver of axial descriptors (stand-off distance, jet envelope), CG governs radial evolution, and PFR exerts moderate influence on radial spread and particle concentration. Cross-system regression shows strong agreement between the two methods (r > 0.8) but with systematic direction-dependent offsets, while a one-way ANOVA confirms statistically significant differences between lateral and coaxial descriptor values (p < 0.05), against a background of low within-configuration repeatability (coefficient of variation below 6%). These findings establish that the two perspectives introduce systematic and configuration-dependent offsets rooted in their optical geometry rather than measurement noise, with each viewpoint capturing complementary information about the PGJS.

This work provides a foundation for standardised PGJS quality assurance in DED-LBp, supporting repeatable and controlled process conditions. The dataset generated is also being used to explore data-driven approaches capable of revealing relationships beyond the reach of the statistical tools applied in this study.