Unveiling the Dual-Coupling Mechanism of Pipe Bursts: Integrating Transient Hydraulics, Geometric Stress, and Adaptive Parallel Computing

Pipe burst risk assessment is frequently hindered by the inability to capture nonlinear interactions between transient hydraulics and complex network geometries. This disconnect often leaves failures in ostensibly safe low-pressure zones unexplained. This study proposes a high-fidelity Computational-Geometric-Physical framework to resolve this challenge. First, to overcome computational bottlenecks, we present a Topology-Adaptive Spectral Partitioning and Geometric Extraction (TAS-PGE) that accelerates large-scale simulations by 1.51×. This efficiency enables our core innovation: the Geometric-Hydraulic Dual Coupling Model (GH-DCSIM). It reveals a critical mechanism where transient waves and acute bends amplify stress far beyond static predictions. Crucially, we prove that traditional static methods underestimate risks in U-bends by 2.9-fold (Safety Factor 2.00 vs 0.68), misidentifying imminent failures as safe. Finally, a variable wave speed mechanism incorporating gas–liquid interaction eliminates physical distortions. To establish physical applicability, we constructed a comprehensive Literature Benchmark Database, comprising real-world accident reconstructions, design limit states, and synthetic parametric scenarios. Independent validation on 34 real-world cases confirms the framework’s robustness, with the mean burst-risk ratio significantly exceeding the failure limit of 1.0 (mean R = 2.80; 95% CI: [1.847, 3.911]). This work shifts the paradigm from pressure-dominated assessment to multiphysics coupling, providing a rigorous basis for predictive maintenance.