Investigating Correlation Between Tensile and Fatigue Strength in DMLM Additive Manufacturing

Conference Abstract: This study investigated potential correlation between tensile properties and fatigue properties across a range of additive manufacturing process conditions. The Direct Metal Laser Melting (DMLM) additive manufacturing process has traditionally prioritized achieving the highest tensile strength. While tensile testing is a quick and straightforward method, it doesn't accurately replicate the complex dynamic failure modes seen in many final product applications where fatigue properties are of critical importance. An initial review of the literature led to the hypothesis that optimal fatigue and tensile properties might not coincide with the same manufacturing conditions [Filippo Berto, A. d. P., Fatigue in Additive Manufactured Metals. 2023, Elsevier].

Four-point bend fatigue properties were evaluated at 15Hz and an R value of 0.1 utilizing both as-printed and as-machined surface roughness samples. S-N curves were generated for each set of manufacturing conditions and utilized to determine the 95% confidence interval lower bound of load capable of achieving at least 10 million cycles. Additional properties such as surface roughness and chemistry were also examined.

Samples were produced on a Renishaw AM500Q utilizing AP&C Grade 23 Ti6Al4V powder conforming to ASTM F3001 and printed at a 30-micron layer height. Laser power (150, 170, 190W), anneal temperature (800, 850, 900C), and anneal cycle count (single/double) were the processing parameters chosen for investigation.

Data revealed that the manufacturing conditions yielding the lowest tensile strength (190W, 900C, double) differed from those resulting in the lowest fatigue strength (190W/800C/single for as-printed, and 150W/800C/single for as-machined). A strong statistical correlation wasn't found between tensile and fatigue properties, however the respective performances were at nearly opposite ends of the DOE. Furthermore, the as-machined fatigue specimens showed more differentiation between process conditions (275MPa range), vs the as-printed fatigue (13MPa range). This finding holds potential implications for selecting appropriate manufacturing conditions for worst-case product performance testing.