Identification of Critical Fault Propagation Paths in Machining Centers Based on Data Envelopment Analysis

Critical fault propagation path identification is essential for fault diagnosis and maintenance decision‐making in machining centers. Existing methods often rely on a single fault‐association representation and rarely evaluate propagation capability together with propagation consequences. This study proposes a data envelopment analysis (DEA)‐based framework for comprehensive evaluation and identification of critical fault propagation paths. A directed fault propagation graph is first constructed by integrating system functional relationships, maintenance experience, and field fault records through FMEA‐based cause–effect analysis, and source‐to‐sink complete paths are extracted using depth‐first search. Edge fault associations are then characterized from three perspectives: statistical dependence, system failure contribution, and conditional propagation. Source‐node failure probability, edge fault association degree, and network topology are fused by the Choquet integral to obtain multi‐representation path fault propagation strengths. Path consequences are evaluated in terms of economic loss and maintenance complexity and transformed into consequence safety margins. Finally, candidate paths are treated as decision‐making units in an output‐oriented CCR‐DEA model, with consequence safety margins as inputs and multi‐representation propagation strengths as outputs. A machining‐center case study, robustness analysis, and comparative validation show that the proposed method can preserve differentiated propagation information while jointly considering propagation capability and consequences, providing a robust basis for critical‐path identification, maintenance prioritization, and resource allocation.