Source Code and Verification Dataset for Automated Staged FE Modeling and Eccentric Prestress Coupling in Railway Continuous Box Girders

Overview This repository provides the computational code, numerical input configurations, and verification datasets supporting the research entitled "Automated Staged Finite Element Modeling and Eccentric Prestress Coupling Verification for Railway Continuous Box Girders." The study develops a physics-consistent automated staged beam–truss finite element framework for railway prestressed concrete continuous box girders. The framework distinguishes numerical solvability, model credibility, and candidate acceptability, with particular emphasis on eccentric prestress transfer, construction-stage state management, and structural and geometric constraint diagnostics. Computational Case The reference case is a three-span (32 + 48 + 32) m railway prestressed concrete continuous box girder. The numerical model is implemented using Python and OpenSeesPy and contains 5,289 nodes, 5,144 finite elements, and 116 registered longitudinal tendon paths. Verification Scope The supporting materials cover the following computational assessments: Analytical P–Pe verification of eccentric tendon–girder force and moment transfer. Full-model tendon–girder connection registration and topology audits. Bridge-scale comparisons between equalDOF_legacy and rigidLink-beam coupling formulations. Prestress-input sensitivity analyses of the rigidLink-beam formulation using scaling factors of 0.8, 0.9, 1.0, 1.1, and 1.2. An independently assembled two-dimensional Euler–Bernoulli benchmark for the representative ZK external-load increment. Construction-stage response auditing and geometric constructability screening. Geometric checking-point refinement and archived candidate-feedback diagnostics. Reproducibility and Limitations The reference numerical analyses adopt small-deformation, linear-elastic modeling assumptions. The prestress-input sensitivity analyses evaluate the rigidLink-beam formulation and do not independently establish the absolute accuracy of the three-dimensional prestress-induced force state. The independent two-dimensional benchmark evaluates the representative external-load increment without tendon prestress. The geometric screening is based on simplified section geometry and nominal duct parameters. The archived exploratory feedback case and the formal reference analysis use different loading and tendon-specification configurations and are reported separately. Users should consult the accompanying README, input files, computational scripts, software environment specifications, and verification records for the scope and execution of the archived calculations. The frozen reference archive contains 12 analysis stages and 294 stage-specific files, including records for 116 tendon bundles and 5,028 prestress-related rigidLink-beam connections. Certain archived scripts retain local Windows path defaults, and independent end-to-end reproduction on a third-party system has not been demonstrated. The historical feedback records concern a separate 100 kN/m configuration and contain documented differences between the original history and derived reviewer trace for iterations 2–6.