Research On Dynamic Deformation Monitoring Of Bridge Structures Based On GNSS And IEWT Techniques

In the last few years,Global Navigation Satellite System(GNSS)technology has been increasingly widely applied in bridge health monitoring(BHM)with the development of the GNSS technology,and the monitoring modes has evolved from long-term deformation monitoring to dynamic deformation monitoring.Due to the influence of multipath effect,instrument background noise,and external environmental noise,the deformation monitoring data from sensors are polluted by noise.It is difficult to directly analyze the bridge GNSS deformation monitoring data.Therefore,the structure modal parameters could be extracted from the monitoring data accurately on the basis of further processing of noise reduction.Based on the GNSS deformation monitoring data of Wilford Suspension Bridge in the UK,this paper focuses on GNSS monitoring data noise reduction,feature extraction,and modal parameter identification centering around Empirical Wavelet Transform(EWT).The main works are as follows:(1)In view of the inaccurate frequency band division and false modes in the traditional EWT processing the GNSS deformation monitoring data,this paper proposed an improved EWT(IEWT)noise reduction method based on the autoregressive power spectral density.Firstly,the autoregressive power spectral density,calculated by the modified covariance method,was used to replace the Fourier spectrum for frequency band division.Secondly,the effective IMF judgment criterion was adopted to extract meaningful components for signal reconstruction.To demonstrate the effectiveness of the IEWT method,a simulated signal with additive noise and field bridge experiments were adopted.The results show that the IEWT method could reduce GNSS measurement noise effectively.Besides,the noise reduction performance of the proposed method was better than that of the Wavelet Transform(WT)and Empirical Mode Decomposition(EMD)methods.(2)A combined method(IEWT-Robust ICA)of IEWT and Robust ICA was put forward to further suppress noise in the effective component frequency bands.Firstly,the multi-mode GNSS bridge deformation monitoring data were denoised and reconstructed by IEWT.Secondly,the obtained reconstructed data,as multi-channel observation signal,were input of Robust ICA.Then,the correlation coefficient and the minimum distortion criterion(MDP)were used to solve the sorting uncertainty,the phase uncertainty,and the amplitude uncertainty of independent components.Finally,the characteristic components of the source signal were extracted.The results show that the noise level of GNSS deformation monitoring data was significantly weakened through the combined IEWT-Robust ICA method.Meanwhile,the difference of maximum dynamic displacement between GNSS dynamic displacement derived by the combined method and the acceleration dynamic displacement obtained by double integration was 0.81 mm.The standard deviation of displacement between GNSS and accelerometer was 0.98 mm,which is less than 1/20 of the bridge allowable deformation displacement.The accuracy is sufficient enough for bridge dynamic deformation monitoring.(3)A modal parameter identification method of Natural Excitation Technique(NEx T),Normalized Hilbert Transform(NHT),and Exponential Decay Function(EDF)was proposed.Based on the structure deformation monitoring data under excitation,this hybrid processing method first separated the characteristic components using the combined IEWT-Robust ICA method.Then,the free decay response of the characteristic component was obtained by NEx T.Moreover,NHT and EDF were applied to identify the vibration frequency as well as damping ratio of mono component.Finally,this method was applied to modal parameter identification of Wilford Suspension Bridge.The fundamental frequency(1.6619 Hz)and damping ratio(0.86%)were detected using this proposed method from the GNSS measurements,which were consistent with the accelerometer results.