Structural identification of the skybridge lifting process for high-rise twin-tower connected structures

To ensure the safety and stability of the high-rise twin-tower connected structure during the integral lifting construction process of the multi-story arched skybridge, this study proposes an early warning framework based on the structural identification (St-Id) theory that combines intelligent algorithms with the finite element (FE) method. Through continuous real-time monitoring of the strain of the skybridge and tower during the lifting process of the skybridge, combined with the environmental vibration tests of the tower, the established FE models of the skybridge and tower are updated, thus realizing the early warning of the lifting process of the multi-story skybridge. For the skybridge, firstly, the structural strain response and temperature data were obtained by instrumenting strain gauges on the skybridge and the towers. To obtain the true strains of the structure, this paper proposed a novel back-propagation neural network optimized by the dung beetle optimizer (DBO-BPNN) to eliminate the effect of temperature, in which the DBO was used to obtain the optimal hyperparameters of the BPNN to avoid the problem that the optimal hyperparameters could not be obtained by the traditional approach (manual empirical selection). Subsequently, the DBO was used to iteratively calibrate the established FE model of the skybridge using the true strain at the time of the skybridge lifting to achieve the accurate prediction and early warning of the subsequent construction process of the skybridge with a prediction error of 10.03 %, proving that the model was able to accurately reflect the real physical state of the structure. For the tower, firstly, the FE model of the super high-rise building was established to analyze the internal force of the tower during the lifting process of the skybridge in combination with the strain monitoring results. Additionally, the modal parameters of the tower before and after the installation of the skybridge were obtained through ambient vibration tests and the effect of lateral support (skybridge) on the tower was analyzed. Then, the calibration and optimization of the FE model of the tower were achieved by considering the influence of the infill wall, and the error of the calibrated FE model was below 6.32 %, thus achieving the safety warning of the tower during the lifting process of the skybridge.