【摘要】 为了更准确地预测和评估钢-混结合段的疲劳性能，设计了缩尺比为1:2的关键格室构件进行设计寿命期内疲劳验证试验，用试验结果验证有限元模型的准确性。采用ABAQUS有限元软件对各种参数下的疲劳应力幅进行计算，结合Eurocode 3中的相关规定预测钢-混结合段的疲劳性能。另外，选择支持向量机对多参数下的钢-混结合段疲劳性能进行评估。采用交叉验证等方法，调优支持向量机的核函数系数G和正则化参数C，以确保模型的最佳性能。研究结果表明：疲劳寿命预测结果准确率达98.78%，该方法为钢-混结合段的疲劳性能研究提供了一种新的、可靠的分析方法，可为工程实际应用提供参考。更多还原

【Abstract】 In order to more accurately predict and evaluate the fatigue performance of the steel-concrete joint section, key cell components on a 1?2 scale for fatigue verification tests were designed within the design life cycle.The experimental results were used to validate the accuracy of the finite element model. Fatigue stress amplitudes with various parameters were calculated using the ABAQUS finite element software, and the fatigue performance of the steel-concrete joint section was predicted based on fatigue details specified in European Specification 3.Additionally, the support vector machine, a powerful machine learning method, was selected to evaluate the fatigue performance of the steel-concrete joint section with multiple parameters. Methods such as Cross-validation are used to tune the kernel function coefficient G and regularization parameter C of the support vector machine to ensure the best performance of the model. The results show that the prediction accuracy of fatigue life is 98.78%. These findings provide a novel and reliable analytical method for studying the fatigue performance of steel-concrete joint sections and offer reference for practical engineering applications.更多还原