Risk Assessment of Long-Span Steel-Truss-Girder Cable-Stayed Bridge with Single Cable Plane and Single Pylon

Long-span single-cable-plane rail-cum-road steel-truss-girder cable-stayed bridges are characterized by complex construction conditions, coupled loading effects, and uncertainties in construction parameters. Conventional risk assessment methods may be limited in quantitative accuracy and computational efficiency, particularly for pre-construction probabilistic prediction. Taking the Qiansimen Jialing River Bridge in Chongqing as the engineering background, this study integrates the analytic hierarchy process (AHP), grey relational analysis, finite element simulation, a radial basis function (RBF) neural network, and Monte Carlo simulation to establish an integrated probabilistic risk assessment framework for bridge construction. Numerical results show that the comprehensive displacement failure probability of the bridge at the maximum cantilever construction stage is 0.12%, indicating a low probability of displacement-related failure under the considered conditions. Cable elastic modulus and initial strain are the dominant parameters affecting cantilever alignment. A 6% deviation from their mean values increases the cantilever displacement by 5.93% and 4.89%, respectively, whereas cable density and Poisson’s ratio have limited effects on structural deformation. The proposed multi-algorithm framework improves the computational efficiency of stochastic probabilistic simulation and provides a quantitative approach for parameter sensitivity analysis and pre-construction risk assessment of this bridge type. The results provide a reference for construction risk control, parameter regulation, and construction scheme optimization for similar urban cross-river rail-cum-road cable-stayed bridges.