Revealing the underlying mechanism in controlling Young’s modulus of additively manufactured Ti-6Al-4V using fuzzified machine learning

Metal products fabricated by additive manufacturing (AM), face challenges in mechanical property evaluation due to the complex thermal history of the process. This study used Ti-6Al-4V alloy, widely used in AM to examine the applicability of the Gibson-Ashby model for samples with porosities lower than 5%. Observations from electron backscatter diffraction (EBSD) and X-ray diffraction (XRD) demonstrated that the Gibson-Ashby model's accuracy is limited due to the changes in the topology of pores and the phase proportions from different processing parameters. However, adaptive neuro fuzzy inference systems (ANFIS) reduced prediction errors on Young's modulus to 0.66 GPa and quantified the combined influence of microstructure variations. The proposed deviation factor addressed the model's neglect of microstructural changes. This laid the foundation for the establishment of a database to precisely control the mechanical properties of the products, thus promoting the optimisation of AM-produced titanium alloy for further practical applications.