TY - BOOK

T1 - Robust Deformation Monitoring of Bridge Structures Using MEMS Accelerometers and Image-Assisted Total Stations

AU - Omidalizarandi, Mohammad

N1 - Doctoral thesis

PY - 2020

Y1 - 2020

N2 - Today, short- and long-term structural health monitoring (SHM) of bridge structures has received considerable attention. However, permanent, cost-effective, and reliable monitoring are still challenging issues. From a surveying or civil engineer's point of view, vibration-based SHM is often carried out by inspecting the changes in the dynamic responses of bridge structures known as modal parameters, such as eigenfrequencies, eigenforms and modal damping. The use of cost-effective micro-electro-mechanical-systems (MEMS) accelerometers with a high sampling frequency is becoming more affordable and feasible for the aforementioned monitoring task. Within this dissertation, a three-step scenario is proposed to choose a suitable MEMS accelerometer despite of its purchase price, measurement range and sampling frequency. Firstly, a robust calibration procedure is proposed and implemented to model MEMS related systematic errors such as biases, scale factors, and non-orthogonality angles between the axes. Secondly, a controlled excitation experiment is conducted by using a high-precision shaker. Thirdly, a static test experiment is accomplished over a long period. Robust, accurate, and automatic estimation of the modal parameters is particularly challenging when vibration measurements are contaminated with a high coloured measurement noise, e.g., due to cost-effective MEMS acceleration data. This is even more challenging when the structure is continuously under imposed forces due to moving vehicles or wind. For this purpose, a robust and automatic vibration analysis procedure the so--called robust time domain modal parameter identification (RT-MPI) approach is proposed and implemented. It is a novel approach in the sense of automatic excitation (e.g. ambient) window selection, automatic and reliable identification of initial eigenfrequencies even closely spaced ones as well as robustly and accurately estimating the modal parameters. To estimate frequencies, damping ratio coefficients, amplitudes, and phase shifts, an observation model consisting of a damped harmonic oscillation (DHO) model, an autoregressive model of coloured measurement noise and a stochastic model in the form of the heavy-tailed family of scaled t-distributions with unknown degree of freedom and scale factor, is employed. The aforementioned three parametric models are jointly adjusted by means of a generalised expectation maximisation (GEM) algorithm. The proposed RT-MPI algorithm is also able to estimate amplitudes in a metric unit and with a high accuracy for the recorded acceleration data by means of double integration of the DHO model. The eigenforms are characterised in a subsequent step, and by using the estimated parameters from the GEM algorithm. In addition, having amplitudes in the metric unit allows to characterise deflection eigenforms in their true scales for selected excitation windows within short time intervals. The deformation/displacement monitoring by merely using the MEMS accelerometer is challenging, since it suffers from accuracy degradation with time for absolute position/displacement estimates. Therefore, the MEMS accelerometers and an image-assisted total station (IATS) are fused by performing one-dimensional (1D) coordinate update within the Kalman filtering framework. To generate 1D displacement data from the IATS, video frames of a passive target, that is attached to a bridge structure, are captured by means of a telescope camera of the IATS. A passive target centroid detection algorithm is proposed and implemented, which is robust and reliable with respect to poor environmental conditions, such as low lighting, dusty situations, and skewed angle targets. Next, an angular conversion factor of the telescope camera is calibrated, which allows to convert the generated displacement data from pixel to metric unit. Experiments are performed in four case studies including simulation, controlled excitation and two real applications of a footbridge structure and a synthetic bridge. The estimated modal parameters are compared and validated by their true values as well as their corresponding estimates obtained from reference sensors such as reference accelerometer, geophone, and laser tracker. Additionally, the estimated eigenfrequencies and damping ratio coefficients are compared with a well-known covariance driven stochastic subspace identification (SSI-COV) approach. The results show that the MEMS accelerometers are suitable for identifying all occurring eigenfrequencies of the bridge structures. Moreover, the vibration analysis procedure demonstrates that amplitudes are estimated in submillimetre range accuracy, frequencies with an accuracy of better than 0.1 Hz and damping ratio coefficients with an accuracy of better than 0.1 and 0.2 % for modal and system damping, respectively. The analysis reveals the superiority of the proposed RT-MPI algorithm compared to the SSI-COV algorithm. Finally, a high accurate displacement time series at the level of submillimetre is generated by fusion of the IATS and the MEMS measurements.

AB - Today, short- and long-term structural health monitoring (SHM) of bridge structures has received considerable attention. However, permanent, cost-effective, and reliable monitoring are still challenging issues. From a surveying or civil engineer's point of view, vibration-based SHM is often carried out by inspecting the changes in the dynamic responses of bridge structures known as modal parameters, such as eigenfrequencies, eigenforms and modal damping. The use of cost-effective micro-electro-mechanical-systems (MEMS) accelerometers with a high sampling frequency is becoming more affordable and feasible for the aforementioned monitoring task. Within this dissertation, a three-step scenario is proposed to choose a suitable MEMS accelerometer despite of its purchase price, measurement range and sampling frequency. Firstly, a robust calibration procedure is proposed and implemented to model MEMS related systematic errors such as biases, scale factors, and non-orthogonality angles between the axes. Secondly, a controlled excitation experiment is conducted by using a high-precision shaker. Thirdly, a static test experiment is accomplished over a long period. Robust, accurate, and automatic estimation of the modal parameters is particularly challenging when vibration measurements are contaminated with a high coloured measurement noise, e.g., due to cost-effective MEMS acceleration data. This is even more challenging when the structure is continuously under imposed forces due to moving vehicles or wind. For this purpose, a robust and automatic vibration analysis procedure the so--called robust time domain modal parameter identification (RT-MPI) approach is proposed and implemented. It is a novel approach in the sense of automatic excitation (e.g. ambient) window selection, automatic and reliable identification of initial eigenfrequencies even closely spaced ones as well as robustly and accurately estimating the modal parameters. To estimate frequencies, damping ratio coefficients, amplitudes, and phase shifts, an observation model consisting of a damped harmonic oscillation (DHO) model, an autoregressive model of coloured measurement noise and a stochastic model in the form of the heavy-tailed family of scaled t-distributions with unknown degree of freedom and scale factor, is employed. The aforementioned three parametric models are jointly adjusted by means of a generalised expectation maximisation (GEM) algorithm. The proposed RT-MPI algorithm is also able to estimate amplitudes in a metric unit and with a high accuracy for the recorded acceleration data by means of double integration of the DHO model. The eigenforms are characterised in a subsequent step, and by using the estimated parameters from the GEM algorithm. In addition, having amplitudes in the metric unit allows to characterise deflection eigenforms in their true scales for selected excitation windows within short time intervals. The deformation/displacement monitoring by merely using the MEMS accelerometer is challenging, since it suffers from accuracy degradation with time for absolute position/displacement estimates. Therefore, the MEMS accelerometers and an image-assisted total station (IATS) are fused by performing one-dimensional (1D) coordinate update within the Kalman filtering framework. To generate 1D displacement data from the IATS, video frames of a passive target, that is attached to a bridge structure, are captured by means of a telescope camera of the IATS. A passive target centroid detection algorithm is proposed and implemented, which is robust and reliable with respect to poor environmental conditions, such as low lighting, dusty situations, and skewed angle targets. Next, an angular conversion factor of the telescope camera is calibrated, which allows to convert the generated displacement data from pixel to metric unit. Experiments are performed in four case studies including simulation, controlled excitation and two real applications of a footbridge structure and a synthetic bridge. The estimated modal parameters are compared and validated by their true values as well as their corresponding estimates obtained from reference sensors such as reference accelerometer, geophone, and laser tracker. Additionally, the estimated eigenfrequencies and damping ratio coefficients are compared with a well-known covariance driven stochastic subspace identification (SSI-COV) approach. The results show that the MEMS accelerometers are suitable for identifying all occurring eigenfrequencies of the bridge structures. Moreover, the vibration analysis procedure demonstrates that amplitudes are estimated in submillimetre range accuracy, frequencies with an accuracy of better than 0.1 Hz and damping ratio coefficients with an accuracy of better than 0.1 and 0.2 % for modal and system damping, respectively. The analysis reveals the superiority of the proposed RT-MPI algorithm compared to the SSI-COV algorithm. Finally, a high accurate displacement time series at the level of submillimetre is generated by fusion of the IATS and the MEMS measurements.

KW - Vibration analysis

KW - Deformation analysis

KW - Modal parameters

KW - MEMS accelerometer

KW - Image-assisted total station

KW - Robust parameter estimation

KW - Generalised expectation maximisation algorithm

KW - Kalman filter

KW - Bridge monitoring

M3 - Doctoral thesis

VL - 859

ER -