A stochastic model based on Markov process is proposed to estimate the number of fiber-break points of a unidirectional fiber-reinforced composite. This model relies on the assumption that each state is assigned in the order of the fiber-break point number, and that it transits to two states from one state. The first state of the two is a state causing a composite fracture, and another is a state of survival while being damaged. Mini composites with 2–4 fibers were prepared using carbon fiber and epoxy resin to examine the applicability of the model. Predicted failure probability curves approach experimentally obtained data with an increase in the fiber-break point and ultimately agree with them. The fiber-break points till fracture are estimated with greater numbers than the number of fibers, which elucidates the significance of the experimental results higher than the average fiber strength. The validity of the model was finally discussed while considering the stress concentration around fiber-break point(s).

提出了一种基于马尔可夫过程的随机模型，用于估计单向纤维增强复合材料的纤维断裂点数量。该模型基于以下假设：每个状态均按光纤中断点编号的顺序分配，并且从一个状态转换为两个状态。两者的第一种状态是引起复合材料断裂的状态，另一种是受损时的生存状态。使用碳纤维和环氧树脂制备了具有2-4根纤维的微型复合材料，以检验该模型的适用性。预测的失效概率曲线会随着纤维断裂点的增加而接近通过实验获得的数据，并最终与他们达成一致。估计直至断裂的纤维断裂点的数量大于纤维的数量，这阐明了实验结果的意义高于平均纤维强度。最后讨论了模型的有效性，同时考虑了纤维断裂点附近的应力集中。