对交通网络级联失效的传播路径进行动态分析是网络脆弱性研究的重要挑战。本文创新性地提出了基于改进耦合映像格子（CML）模型和Gephi动态图分析方法，能有效地识别级联失效传播的脆弱路径及关键节点。以福建省高速公路网络为案例，首先通过增加隧道耦合因子和流量耦合因子对传统CML模型进行了改进，并获取网络在不同攻击策略下级联失效的全局失效扰动阈值及传播速度。其次，利用Gephi动态图功能捕捉级联失效的动态传播过程，提出“节点失效树状图”来描述节点失效的时序和延续关系。此外，采用逆序标号、正序搜寻的算法逻辑，提出“失效节点逆序标号法”以实现级联失效传播的脆弱路径搜寻，并识别出失效路径中的关键节点。结果表明：虽然所有路网节点均具有一定抵御扰动的能力（全局失效扰动阈值R≤1.1时），但不同节点抵御扰动的能力差异较大（1.2≤R≤4.8），平均全局失效扰动阈值为2.02；选取四类特征点（最大节点度、最大介数、最大V/C比、最大隧道因子）进行攻击，其中基于最大介数和最大V/C比的攻击产生级联失效风险最大，其失效节点占比分别只需17和21个时间步即达到失效高峰；进一步地，进行最脆弱路径和次脆弱路径的关键节点搜寻，共识别出8个关键节点，提高其鲁棒性可显著抑制级联失效的传播规模。本研究能为交通网络脆弱性的动态分析提供新视角，为路网抗风险方案制定提供决策支持。

The dynamic analysis of cascade failure propagation paths in transportation networks remains a notable challenge in network vulnerability research. This study introduces an innovative integrated approach that combines an enhanced coupled map lattice (CML) model with Gephi 's dynamic graph analysis to identify vulnerable cascade paths and critical nodes effectively. Using Fujian Province's expressway network as a case study, the traditional CML model is enhanced by incorporating tunnel and traffic coupling factors, enabling the determination of trigger thresholds and propagation speeds of cascade failures under various attack strategies. Gephi's dynamic graph function captures the cascade failure propagation process, and we propose a "node failure tree diagram" to illustrate the temporal and sequential continuity relationships of node failures. Additionally, a "reverse-order marking method for failed nodes" is developed using a reverse-order marking and forward-order searching algorithm to identify vulnerable paths and critical nodes in failure propagation. The results demonstrate  that while all nodes exhibit some perturbation resistance (failure threshold R ≤ 1.1), robustness varies significantly (1.2 ≤ R ≤ 4.8), with an average threshold of 2.02. Among four targeted attacks—maximum node degree, betweenness, V/C ratio, and tunnel factor—attacks based on maximum betweenness and V/C ratio are most critical, reaching peak failure in 17 and 21 time-steps, respectively. Identifying critical nodes in primary and secondary vulnerable paths revealed eight key nodes, whose enhance the robustness significantly suppresses the scale of cascade propagation. This study offers a novel approach to analyzing transportation network vulnerability and supports risk-resistant network planning.