Answers are grounded in this paper and cite their sources.
M2 high-speed steel (HSS) exhibits high hardness and wear resistance, but its high carbon and alloy contents make crack-free fabrication by laser-based powder bed fusion (PBF–LB/M) challenging. This study develops a data-driven strategy for processing M2 HSS without preheating the baseplate. An initial dataset of 27 experiments varying laser power, scanning speed, and hatch spacing was used to predict hardness and relative density using five machine-learning algorithms and quadratic response surface modelling. Model optimization and evaluation were performed using repeated nested cross-validation. Random Forest achieved the highest predictive performance for hardness (R2 = 0.592 ± 0.078), while Gradient Boosting Regression performed best for relative density (R2 = 0.619 ± 0.048). Sensitivity analysis identified hatch spacing as the dominant processing parameter. Consequently, hatch spacings of 55–70 µm were explored computationally, and 60 µm was selected for experimental verification. This condition produced a relative density of 99.94–99.99% and an HV hardness of 718–782, while suppressing the delamination observed at larger hatch spacings by increasing melt-track overlap and improving consolidation. Microstructural characterization revealed a refined non-equilibrium structure containing approximately 50.3% retained austenite and 24% bcc phase, with an average grain size of ~ 0.8 µm. Electron backscatter diffraction revealed a strong crystallographic texture (10.21 MRD) tilted 31.3° from the build direction toward [101]. These results demonstrate that data-driven parameter selection, combined with experimental verification, provides an effective strategy for improving the processability of crack-sensitive M2 HSS via PBF–LB/M.
No takes yet. Share an insight, caveat, or question.
Sahu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: