In-situ fatigue crack propagation experiment was conducted on laser cladding with coaxial powder feeding (LCPF) K477 under various stress ratios and temperatures. Multiple crack initiation sites were observed by using in-situ scanning electron microscopy (SEM). The fatigue short crack growth rate was measured, and the impacts of temperature and stress ratio on this growth rate were analyzed. Based on these experiments, the experimental data were expanded, and three ensemble learning algorithms, i.e. random forest (RF), extreme gradient boosting (XGBoost) and light gradient boosting machine (LightGBM) were employed to establish a fatigue short crack growth rate model controlled by multiple parameters. It is indicated that the RF model performs the best, achieving a coefficient of determination ( R 2) of up to 0.88. The fatigue life predicted by the machine learning (ML) method agrees well with the experimental one.

Orthogonal experiment design together with the analysis of variance was used to examine the processing parameters (laser power, scan speed, layer thickness and hatch spacing) of selective laser melting (SLM) for superior properties of SLM parts, in which nine groups of specimens of Ti-6Al-4V were fabricated. The porosity for each group was measured and the results clarify that the influence sequence of individual parameter on the porosity is laser power > hatch spacing > layer thickness > scan speed. Ultrasonic fatigue tests (20 kHz) were conducted for the SLMed specimens in high-cycle fatigue (HCF) and very-high-cycle fatigue (VHCF) regimes. The S-N data show that the fatigue strength is greatly affected by the porosity: the group with the smallest porosity percentage having the highest fatigue strength in HCF and VHCF regimes. Moreover, the observations by scanning electron microscopy revealed that fatigue cracks initiate at lack-of-fusion defects in the cases of surface and internal crack initiation.