Considering the large dispersion when processing small-sample fatigue test data of composite materials, a new method for modeling probabilistic S-N curves is proposed in terms of the equivalent fatigue lives. An equivalent fatigue life conversion model is first established based on two fundamental assumptions to improve the small-sample information utilized. Subsequently, a backward statistical inference technique improved by particle swarm optimization is used to determine probabilistic S-N curves through the equivalent fatigue lives. Finally, the proposed method is verified in terms of precision and stability by the fatigue test data of carbon eight-harness-satin/epoxy laminate. The results indicate that the proposed method can offer an accurate description of the probabilistic behavior of composite materials with small-sample test data.

考虑到大型分散处理的复合材料的小样本疲劳测试数据，用于建模概率的新方法，当小号- ñ曲线在等效疲劳寿命方面提出。首先基于两个基本假设建立等效疲劳寿命转换模型，以改善所利用的小样本信息。随后，通过粒子群优化改进的后向统计推断技术用于确定概率S - N曲线通过等效疲劳寿命。最后，通过碳八线束缎/环氧树脂层压板的疲劳试验数据，从精度和稳定性方面验证了该方法的有效性。结果表明，所提出的方法可以用小样本测试数据准确描述复合材料的概率行为。