الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 166 |  |  |
| | | from | 166 | | |
Abstract
Structural Health Monitoring (SHM) is an emerging field in civil engineering that offers the potential for continuous and periodic assessment of the safety and integrity of civil infrastructures. Based on previous structure condition knowledge, certain preventive measures can be adopted to prolong the service life of the structure and prevent catastrophic failure. Global Positioning System (GPS) is one of the major methods utilized in SHM. However, as the technical feasibility of using GPS for recording relative displacements has been proven, the challenge issue for users is to determine how to make use of the relative displacements being recorded. This thesis proposes a mathematical framework that supports the use of Real Time Kinematics (RTK)-GPS data for structural monitoring. The used GPS data is acquired by the Research Center of Structural Health Monitoring and Control of Harbin Institute of Technology (HIT), to monitor the loads and responses of one cable-stayed bridge. After one year of reconstruction of this bridge, cracks were observed at 48.2 m from the first abutment again. The analysis will include the movements and damages of the bridge, the current applied operational safety methods, and the bridge cracks cause under different stress factors such as wind speed, temperature change and traffic loads. In addition, the bridge towers lateral, longitudinal movements are to be analyzed Signal processing methods (Kalman and Adaptive filtering, parametric least square methods were utilized to smooth and denoise the recorded GPS signals) and (Wavelet analysis algorithm (STFT,CWT,DWT) are to be used to transform the GPS signals into the frequency time domain and detect different bridge movements and possible damage). The geometrical analysis methods (plane model, span length model and polar coordinates model) are to be used for the bridge movements and damages analysis. For the analysis of the periodic components of the GPS signals, a high-pass filtering process is applied to the signal to determine the high frequencies components of the bridge movements. The transformation of the recorded signals from the time domain to the frequency domain are formed by power spectrum Dissertation for Doctoral Degree in Engineering, Harbin Institute of Technology -IVdensity (PSD) and short time Fast Fourier Transformation (STFT), then the power spectrums of the transformed signals are analyzed. The following are some of the analysis the results: (1) The traffic loads are the main factor that affects the bridge damage, (2) The maximum deformation was pronounced at 48.2 m from the first abutment six months after the bridge use, (3) The STFT is a significant step forward from the traditional FFT in terms of structural response analysis, (4) The sensitivity of the recorded GPS signals does not depend on the position of the GPS sensors, (5) The geometric analysis method provides an easy way to calculate the bridge tower movement, whereas, the physical analysis method is better for detecting the bridge damage, (6) The observed frequency using a 20 HZ GPS is not suitable for monitoring the observed structure natural frequency, (7) The Kalman Filtering is suitable for the dynamic study, (8) It was found that the PSD and STFT analysis results reflect the bridge expected movements and cracks, and (9) It was found that the bridge cracks are mainly caused by the shear force due to movement of the south tower and the bridge nonlinear movements due to traffic loads. GPS and accelerometer techniques are used to collect the lateral displacements, acceleration and torsion displacements data of the bridge tower. Analysis the collected data provide the following results: (1) The recorded GPS signals are noisy, (2) The Wden function provides a 20% increase of accuracy, (3) Power spectral density is a good parameter to detect the tower movements, and (4) GPS can be used as a trustworthy tool for characterizing the dynamic behavior of the low frequency bridges. In addition to previous analysis methods, two identification models namely; Multi Input-Single Output (MISO) robust fit regression and Neural Network Auto- Regression Moving Average with eXogenous input (NNARMAX) models are used for the identification of the bridge movements. The analysis of test results indicate that: (1) The NNARMAX [4411] and [5415] models defined by taking into account the results of robust regression analysis estimate structural movements are more accurate than the NNARMAX[0100] model, and (2) The robust fit regression models have good capacities for mapping the relationship between the applied loads effects factors and the tower displacements. Temperature and humidity effects on the entire modals shapes are insignificant. Abstract - V - Keywords: Global Position System; Structural Health Monitoring; Bridge; Kalman Filter; Parametric Least Square method; Wavelet Transformation; Neural Network Auto-Regression Moving Average with eXogenous; Monitoring.