【摘要】 以气缸盖铸铁材料为研究对象，采用本体取样的方式进行了一系列不同温度范围下的热机械疲劳试验。结果表明，铸铁材料的疲劳试验表现出循环软化、循环稳定和快速失效3个阶段。此外，反相位加载下材料的疲劳寿命显著小于正相位加载。使用人工神经网络（Artificial Neural Network, ANN）、随机森林（Random Forest, RF）等6种典型有监督学习模型对试验数据进行疲劳寿命预测；结果表明，模型无法学习到材料的疲劳寿命分布趋势。针对该问题，利用基于生成对抗网络（Generative Adversarial Network, GAN）自监督算法，实现了对气缸盖铸铁材料热机械疲劳寿命的预测，在小样本条件下表现出了较好的预测效果。该研究对于开展气缸盖设计和疲劳分析有着极强的指导意义和参考价值。更多还原

【Abstract】 Taking cast iron material of cylinder head as the research object, a series of thermo-mechanical fatigue experiments under different temperature ranges were conducted through bulk sampling. The results show that the fatigue test of cast iron materials exhibits three stages: cyclic softening, cyclic stability and rapid failure. Additionally, the fatigue life of materials under inverse phase loading is significantly shorter than that under positive phase loading. Six typical supervised learning models, including artificial neural networks(ANN) and random forest(RF), were used to predict the fatigue life of the experimental data. However, the results indicate that these models failed to learn the fatigue life distribution trend of the materials. For this problem, the prediction of the thermal mechanical fatigue life of cast iron materials for cylinder heads was achieved by using the self-supervised algorithm based on the generative adversarial network(GAN), and it showed a good prediction effect under the condition of small samples. This research has strong guiding significance and reference value for cylinder head design and fatigue analysis.更多还原