Fatigue Life Prediction and Optimization of AA2219 Using Taguchi and Regression Analysis

In the present study, the fatigue life of aerospace-grade AA2219 aluminium alloy was systematically evaluated under varying cyclic loading conditions using a combined Taguchi experimental design and regression analysis. AA2219, widely deployed in cryogenic tanks and structural aerospace components, exhibits sensitivity to fatigue damage due to its precipitation-hardened microstructure. A two-factor, three-level Taguchi L9 orthogonal array was implemented, considering maximum stress level and loading frequency as critical input parameters. Experimental fatigue tests were conducted per ASTM E466, and the resulting fatigue lives were analysed through ANOVA to identify dominant influencing factors. The maximum stress level emerged as the most significant factor, contributing 6.83% to fatigue life variation, while loading frequency exhibited a secondary effect. A second-order polynomial regression model was developed and validated, exhibiting excellent accuracy with an R² of 0.9997 and RMSE of 183.86. SEM-based fractography confirmed distinct fatigue mechanisms under low and high stress regimes, including striation development and brittle fracture morphologies. A confirmation experiment demonstrated only 2.15% deviation between predicted and experimental fatigue life, validating the model's reliability. The proposed methodology offers a predictive and optimization framework to enhance the fatigue performance of AA2219, reducing experimental workload while providing actionable insights for aerospace structural design. Major Findings: Maximum stress level emerged as the key parameter influencing fatigue life, contributing 6.83% to the total variation observed. A robust second-order regression model (R² = 0.9997) was developed, predicting fatigue life with only a 2.15% deviation from actual experiments, proving its reliability. SEM-based fractographic analysis revealed a clear transition: fine fatigue striations dominated at low stress, while brittle, quasi-cleavage features took over at higher stresses.