关键词: 损伤力学, 支持向量机, 铝合金, 疲劳, 寿命预测

Abstract: Based on the continuous damage mechanics (CDM) theory and Support Vector Machine (SVM) model, a novel fa-tigue life prediction method has been developed, aiming to improve the accuracy of fatigue life prediction for 2014-T6 aluminum alloy materials. Firstly, by adopting the continuous damage mechanics model and the secondary develop-ment of the Abaqus-based UMAT subroutine, a damage mechanics finite element numerical implementation method for predicting the fatigue life of 2014-T6 aluminum alloy was established, and a calibration method for material pa-rameters based on the Particle Swarm Optimization algorithm was proposed. Subsequently, to further optimize the prediction results, the SVM model was employed to train the errors of fatigue life prediction results based on damage mechanics, thereby correcting the numerical prediction results. By comparing the prediction results of damage me-chanics finite element, the SVM model-corrected prediction results, and experimental results, it was found that the accuracy of the SVM model-corrected prediction results was higher, verifying the applicability of the proposed method. This provides an effective solution for the fatigue life prediction of 2014-T6 aluminum alloy and is expected to play a significant role in practical engineering applications.

Key words: Damage mechanics, Support vector machine, Aluminum alloy, Fatigue, Life prediction