Thermomechanical coupling mechanism and prediction model of flexible composite pipes under the combined action of internal and external temperatures

Given that real-time monitoring of pipe-wall stress in buried flexible composite pipes is challenging, while stress is strongly coupled to both ground-surface temperature and crude-oil temperature, this study clarifies the thermomechanical coupling effects and mechanisms of buried flexible composite pipes under extreme internal–external temperature fields, combined internal-pressure/soil-pressure stress fields, and pipe–soil interaction. A multi-parameter, robust machine-learning prediction model is developed. Using a flexible composite pipe from a representative oil-and-gas gathering and transportation system as the target, a COMSOL multiphysics model for crude-oil transport is established, incorporating modules for normal operation, isothermal holding, and internal-fluid-heat-source-driven cooling. An orthogonal simulation design covering seven ground-surface temperatures, five internal pressures, and three burial depths is conducted to obtain time-series distributions of temperature and stress for each pipe-wall layer. In addition, a thermal cycling loading platform for a buried pipe section is built, enabling synchronized acquisition of interlayer temperature-sensor and stress-sensor data to obtain complete time-history responses. By fusing simulation outputs with experimental measurements, more than 10,000 σT (temperature-stress product) samples are generated, and a thermomechanical prediction model suitable for extreme environments is trained. The results indicate that temperature and stress responses exhibit similar evolutionary patterns; the interaction between ground-surface temperature and internal temperature causes a rapid increase in interlayer shear stress and triggers stress redistribution; and pressure-induced stress shows a time lag relative to temperature and its gradients. After introducing quantitative descriptors for extreme temperature-difference variations, pressure changes, and heat-transfer delay, high-accuracy prediction is achieved. This framework provides a theoretical basis for safety assessment and digital monitoring of flexible composite pipes during crude-oil transportation. A graphical abstract is presented to summarize a closed-loop modeling–testing–prediction framework for buried flexible composite pipes operating in an “internally hot–externally cold” environment. A transient COMSOL multiphysics model couples radial heat diffusion, multilayer thermomechanics, internal pressure, and pipe–soil restraint to reproduce the normal operation–isothermal hold–cool-down sequence under varying ground-surface temperature and burial depth, yielding layerwise temperature/stress histories and the evolution of interfacial shear. A buried thermal-cycling rig provides synchronized interlayer temperature and stress measurements for model validation and data fusion. The integrated simulation–experimental dataset yields more than 10,000 σT samples to train a robust machine-learning surrogate, enabling rapid prediction of key thermomechanical indicators and supporting digital monitoring and service-life assessment under extreme conditions.