内聚力模型（Cohesive Zone Model，CZM）因其参数设置简便、数值稳定性高，并能有效模拟复杂裂纹扩展行为，在飞机金属薄壁结构的剩余强度评估中展现出重要价值。本文围绕其牵引-分离定律（Traction-Separation Law， TSL），系统梳理了典型TSL曲线的几何特征、初始刚度及关键参数的物理意义，并对比分析了不同TSL形状在金属延性断裂模拟中的适用性与差异性。在此基础上，综述了内聚参数的实验测量方法与数值反演技术，探讨了参数选取对有限元模拟精度的影响。此外，针对CZM的数值实现方式，本文分类阐述了二维内聚单元、壳单元及三维内聚单元三种建模策略，并对比了各自的优缺点。通过典型工程案例，验证了CZM在金属薄壁结构断裂模拟中的可行性与适用性。展望未来，为进一步提升内聚力模型在金属薄壁结构中的应用水平，亟需解决两大关键科学问题：不同TSL形状下内聚参数的定量关系、TSL参数与微观损伤机理的关联机制。

The Cohesive Zone Model (CZM) demonstrates significant value in the residual strength assessment of aircraft metallic thin-walled structures because of its straightforward parameter setting, high numerical stability, and ability to effectively simulate complex crack propagation behavior. This paper centers on its Traction-Separation Law (TSL), systematically reviewing the geometric characteris-tics of typical TSL curves, initial stiffness, and the physical significance of key parameters. It also compares and analyses the ap-plicability and differences between various TSL forms in simulating ductile fracture of metals. Building on this, the paper summariz-es experimental measurement methods and numerical inversion techniques for cohesive parameters, and explores the influence of parameter selection on the accuracy of finite element simulations. Furthermore, regarding the numerical implementation of CZM, the paper categorizes and explains three modelling strategies: two-dimensional cohesive elements, shell elements, and three-dimensional cohesive elements, comparing their respective advantages and disadvantages. Through typical engineering case studies, the feasibility and applicability of CZM in simulating fracture in metallic thin-walled structures are validated. Looking ahead, to further enhance the application level of the cohesive zone model in metallic thin-walled structures, two critical scientific problems urgently need to be addressed: the quantitative relationship between cohesive parameters under different TSL shapes, and the correlation mechanism between TSL parameters and microscopic damage mechanisms.