Data-driven insights into the dynamic operational performance of large-scaled complex titanium alloy castings

In the rigorous environments of aerospace engines, the microstructure and mechanical properties of service components undergo dynamic and heterogeneous evolution. This presents significant challenges in identifying performance-sensitive parts and key microstructural control indicators. This study addresses this complexity by concentrating on large-scaled complex titanium alloy castings (LCTACs) utilized in aerospace engines, implementing a data-driven approach to microstructural analysis. A grey relational analysis (GRA) model is developed based on detailed dissection and data mining of LCTACs across various service durations, elucidating the evolving microstructure-property relationships during service. Grey relational coefficient heatmaps indicate that performance-sensitive parts of the LCTAC are mainly concentrated in the thicker zones of the outer ring, where the decrease in tensile strength is twice that of the thin-walled inner ring. The ranking and analysis of grey relational degrees demonstrate that, with prolonged service duration, the extreme values in grain size distribution progressively outweigh the expectation of grain size as a critical factor, decisively influencing the tensile strength of the LCTAC.