In this work, a robust method for determining the input fatigue parameters of a fatigue cohesive zone model recently developed by Dávila is presented. The method is based on surrogate modelling and multi objective optimization. Experimental data from delamination tests is used for the parameter determination, highlighting the phenomenological nature of the model. The ability to simulate both fatigue onset and propagation is verified for two different monotonic cohesive laws. The predictive capability of the methodology is validated against the experimental results of a partially reinforced double cantilever beam specimen (R-DCB) with non-self-similar crack propagation.

在这项工作中，提出了一种用于确定 Dávila 最近开发的疲劳粘性区模型的输入疲劳参数的稳健方法。该方法基于代理建模和多目标优化。来自分层测试的实验数据用于参数确定，突出了模型的现象学性质。针对两种不同的单调内聚定律验证了模拟疲劳开始和传播的能力。该方法的预测能力针对具有非自相似裂纹扩展的部分增强双悬臂梁试样 (R-DCB) 的实验结果进行了验证。