Stress-Strain Analysis and Fatigue Life Prediction of SnBiInZn High-Entropy Alloy QFN Solder Joints Under Random Vibration

A finite element model of SnBiInZn high-entropy alloy(HEA) QFN solder joints was established and subjected to modal analysis. A random vibration stress-strain simulation was conducted, and the simulation results were validated experimentally. Based on Manson’s empirical formula for high-cycle fatigue, combined with Miner’s linear cumulative damage theory and the Steinberg model, the fatigue life of different solder joint materials under random vibration loading was comparatively analyzed. Building upon the confirmed significant influence of key structural parameters—namely solder joint height, pad length, and pad width—a regression analysis was performed using orthogonal experimental design combined with machine learning BP/MLP models. A predictive model for the random vibration fatigue life of QFN solder joints was thereby established, enabling effective prediction of fatigue life under arbitrary combinations of structural parameter levels. The results indicate that among the four solder joint materials (SAC305, SAC387, 62Sn36Pb2Ag, and SnBiInZn HEA), SAC387 exhibits the shortest vibration fatigue life, while the SnBiInZn high-entropy alloy demonstrates the optimal fatigue life, approximately 2.3 times that of SAC305. The developed BP/MLP model achieves high prediction accuracy, with the relative error between predicted and simulated values controlled within 5%.