Investigation Of And Modal Parameter Identification Algorithm Of Super High-rise Buildings And Long-span Bridges Based On GNSS Measurement

Modal parameter identification of super high-rise buildings and long-span bridges is a research hotspot in the field of structural health monitoring.It is extremely significant to real-time understand the real change of the modal parameters of structures in ensuring safe operation of structures.The traditional parameter identification algorithms extract modal parameters relying on input and output data.It needs artificial excitation,and is easy to damage the structures.Only based on responses,parameter identification algorithms under environmental excitation identify the structural modal parameters under structural operational conditions.However,the existing algorithms still need further study and improvement in the aspects of calculation accuracy and calculation efficiency.Based on multi-system RTK-GNSS(Real Time Kinematic-Global Navigation Satellite Systems)signals,this paper systematically investigated the modal parameter identification algorithms of structure.The main works are as follows:(1)The distribution characteristics of single system,double system and three system RTK-GNSS background noise were studied.A combined filtering method(CC-CEEMDAN-WP)of complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN),correlation coefficient and wavelet packet(WP)technique was proposed.For the sake of testing the effect of noise reduction via employing CC-CEEMDAN-WP,a simulated signal with additive noise was introduced.The results show that the signal-to-noise ratio(SNR)was increased from1.36 to 10.36 after applying CC-CEEMDAN-WP method.Meanwhile,the normalized root mean square error(NRMSE)was decreased from 14.82% to 8.31%.Finally,the noise was effectively weakened.(2)An improved stochastic subspace algorithm based on data-driven was put forward.This algorithm obtains the R-matrix in QR decomposition via constructing a new matrix F,which not only guarantees the identification accuracy,but also improves the computational efficiency.To demonstrate the effectiveness of these methods,a five-degree-of-freedom interlaminar shear model was adopted.The result shows that the running time of the improved algorithm is 20% of the traditional algorithm.Based on the RTK-GNSS monitoring data,the problem of modal parameter identification of a super high-rise structure(i.e.Tianjin 117 Tower)was investigated by using the proposed algorithm.The modal parameters of the structure were successfully extracted from out-put signals.Meanwhile,to contrast the field measurement results,the three-dimensional finite element model(FEM)of the structure is established by using Ansys finite element analysis software.Though modal analysis,the first six order modal frequencies and the corresponding modes of the structure were obtained.Finally,the result shows that the identified results are in good agreement with the finite element results.(3)Three kinds of multi-sensor multi-rate Extended Kalman Particle Filter(EKPF),Unscented Kalman Particle Filter(UKPF)and Cubature Kalman Particle Filter(CKPF)weighted data fusion algorithms were put forward.These three algorithms have two layers structure.In the first layer,EKPF,UKPF or CKPF were used to obtain the local estimation of each sensor.In the second layer,the local estimation results from different sensors were fused based on scalar linear minimum weighted fusion criterion.Subsequently,the optimal fusion estimation results in the minimum sense of variance were derived.Furthermore,the feasibility of these three algorithms was demonstrated via a three-sensor control system and a five-sensor target tracking system.The analysis results show that the fusion results obtained via above algorithms are in good agreement with the real state of the system,and their performance is superior to that of the estimation results of single sensor.(4)An approach integrating RTK-GNSS and accelerometer sensors to monitor structural dynamic deformation of long-span bridges was developed.The acceleration signals with high sampling rate were fused with GNSS signals with low sampling rate by employing the proposed weighted data fusion algorithms.This solves the problem of low sensitivity of GNSS sensor to high frequency signal.Then,two long-span bridges(i.e.Tianjin Rainbow Bridge and Tianjin Yonghe Bridge)are taken as research objects to verify the effectiveness of the proposed algorithm.The results show that only a small number of low-order modes can be identified based on the GNSS monitoring results,while higher-order modal information can be identified based on the fusion estimation results.Meanwhile,the three-dimensional finite element models of the structures were established before the field measurement,and the first twelve order modal frequencies and modal shapes of the structures were predicted.It can be found that the natural frequencies derived experimentally coincide with the predicted value based on the finite element models.