Quantitative analysis of urban underground drainage pipeline service status under rainstorm disaster

Abstract Underground drainage pipeline (UDP) system is critical infrastructure for urban flood control and drainage. During extreme rainstorms, rainfall infiltration can alter pore-water pressure and soil-pipe interaction in the surrounding soil. These changes may affect the structural safety of buried pipelines under traffic load. To investigate the structural responses of buried pipelines during extreme rainstorms and provide a basis for emergency assessment, this study integrates full-scale mechanical test, finite element simulation, historical rainstorm scenarios, and a one-dimensional convolutional neural network (1D-CNN) data agent model. Firstly, full-scale pipeline mechanical tests are conducted, and test results are used to calibrate material parameters in numerical simulations; on this basis, a pavement-pipeline coupled numerical model is established to evaluate UDP structural responses under normal service state. Secondly, based on time-series rainfall data from historical rainstorm disaster in Beijing, China, 20 numerical working conditions are established to investigate the effects of pipe voids, burial depth, subgrade soil permeability, and pipeline diameter on structural responses under the coupled action of rainstorm and vehicle load. Finally, a dataset containing 2370 samples is used to develop the 1D-CNN model; this model achieves an MSE of 0.0003 and an R² of 0.9955 on the test set. The proposed framework provides structural response prediction and identifies potentially vulnerable locations, offering a quantitative reference for emergency inspection prioritisation and structural safety assessment of urban pavement-pipeline systems during extreme rainstorms.