Data-Driven Sensitivity Analysis for Static Mechanical Properties of Additively Manufactured Ti–6Al–4V

Abstract Additive manufacturing (AM) has been extensively investigated in recent years to explore its application a wide range of engineering functionalities, such as mechanical, acoustic, thermal, and electrical properties. A data-driven approach is proposed investigate the influence major fabrication parameters laser-based additively manufactured Ti–6Al–4V. Two separate powder bed fusion techniques, i.e., selective laser melting (SLM) direct metal sintering (DMLS), have several data regarding tensile properties Ti–6Al–4V alloy with their corresponding are collected from open literature. Statistical analysis performed for four (scanning speed, power, hatch spacing, layer thickness) three postfabrication (heating temperature, heating time, hot isostatically pressed or not) which influencing factors by researchers identify behavior on static mechanical (i.e., yielding strength, ultimate elongation). To relationship between input output parameters, both linear regression artificial neural network (ANN) models developed using 53 100 datasets SLM DMLS processes, respectively. The model resulted an average R squared value 0.351 0.507 compared 0.908 0.833 case nonlinear ANN modeling based modeling, Both local global sensitivity analyses carried out important future optimal design. Based current study, (SA) suggests that most sensitive scanning heat treatment temperature while power. In fabricated alloy, scan speed found be impactful time turned least affecting parameter. results can used tailor alloy's per requirement could useful optimize already tailored design Sobol's implicates spacing influential strength thickness followed prominent elongation alloy. Future work would still needed eradicate some limitations this study related limited dataset availability.