Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/86433
Title: Application of GPS for monitoring long-span cable-supported bridges under high winds
Authors: Chan, Wai-shan
Degree: Ph.D.
Issue Date: 2010
Abstract: Many long-span cable-supported bridges have been built across the world in the past two decades to meet the economic, social, and recreational needs of communities. However, due to the highly flexible nature and low damping levels of such bridges, they are vulnerable to wind-induced vibration. To protect the immense capital investments made in such bridges and ensure user comfort and safety during the service stage, the implementation of long-term structural health monitoring systems (SHMS) on long-span cable-supported bridges has become a trend in wind-prone regions. Among all measurement parameters, displacement is a paramount variable used to assess the serviceability, safety and integrity of long-span cable-supported bridges. However, it is difficult to measure the wind-induced absolute displacement of a long-span cable-supported bridge, which includes both a static component and a dynamic fluctuating component, by using traditional sensors. One alternative solution is to use global positioning systems (GPS), but the application of the GPS for bridge monitoring engenders many challenges to professionals. This thesis therefore focuses on the application and integration of global positioning systems (GPS) with structural health monitoring systems (SHMS) and computer simulation to monitor and assess the serviceability and strength of long-span cable-supported bridges under strong winds. Although calibration works of GPS have been performed by other researchers for building structures, the performance of GPS must be thoroughly validated for application to long-span cable-supported bridges because the fundamental frequency of a long-span cable-supported bridge is often much lower than that of a building. In this connection, a motion simulation table is designed and manufactured as a test station that simulates various types of two-dimensional motions across a wider range of frequencies in either the horizontal plane or the vertical plane. A detailed calibration study is then carried out in an open area in Hong Kong using the motion simulation table to assess the dynamic displacement measurement accuracy of the GPS in the longitudinal, lateral, and vertical directions. In the calibration study, the static tests are first carried out with stationary antennae to identify the background noise in the GPS measurements. An examination of statistical data recorded over a period of 9 hours shows that the background noise is dominated by low frequency components. A band-pass filtering scheme for sinusoidal motion and circular motion is designed and applied to the displacement data recorded by the GPS, which are then compared with those generated by the table. The comparative results show that for two-dimensional sinusoidal and circular motions in the horizontal plane and one-dimensional sinusoidal motions in the vertical direction, the GPS can be used to obtain accurate dynamic displacement measurements if the motion amplitude is no less than 5 mm in the horizontal plane or 10 mm in the vertical direction and the motion frequency is less than or equal to 1 Hz. The dynamic displacement measurement accuracy of the GPS is finally assessed using the measurement data of wind-induced two-dimensional dynamic displacement responses of the Di Wang Tower in the horizontal plane during Typhoon York and wind-induced one-dimensional dynamic displacement response of the Tsing Ma suspension bridge deck in the vertical direction during Typhoon Victor. The comparative results demonstrate that the GPS can trace complex wind-induced dynamic displacement responses of real structures satisfactorily. However, the reduced accuracy of GPS displacement measurements due to multipath effects and the low sampling frequency of the GPS receivers is also highlighted in the motion simulation table tests described above. To enhance the measurement accuracy of the total (static plus dynamic) displacement response of civil engineering structures, the concept of integrating signals from a GPS and an accelerometer for deformation monitoring is suggested. This thesis presents two frameworks of integrated data processing techniques that use both empirical mode decomposition (EMD) and an adaptive filter. To assess the effectiveness of the proposed integrated data processing techniques, a series of motion simulation tests simulating various types of motion around a pre-defined static position are performed at a site and recorded by a GPS receiver and an accelerometer. The proposed data processing techniques are then applied to the recorded GPS and accelerometer data to find both static and dynamic displacements. These results are compared with the actual displacement motion generated by the motion simulation table. The comparative results demonstrate that the proposed technique can significantly enhance total displacement measurement accuracy. Wind and structural health monitoring systems including GPS have been installed on some long-span cable-supported bridges, but it is not clear how to use GPS data to assess the serviceability and strength of the bridge under strong winds. This thesis takes the Tsing Ma Bridge as an example to manifest how the GPS and anemometers installed on the bridge can be used for this purpose. The Tsing Ma Bridge in Hong Kong is a long suspension bridge, and a Wind And Structural Health Monitoring System (WASHMS) that includes 6 anemometers and 14 GPS stations has been fully operational on the Tsing Ma Bridge since 1997 (for the anemometers) and 2002 (for the GPS stations), respectively. Because Hong Kong is situated in an active typhoon region and encompassed by a complicated topography, wind characteristics around the Tsing Ma Bridge are very complicated. The wind environment surrounding the Tsing Ma Bridge is thus ascertained by analyzing long-term wind data recorded by the anemometers for both typhoons and strong monsoons. The wind measurement data taken in the field are first pre-processed in an attempt to produce a high quality database. The wind measurement data recorded by the ultrasonic anemometers in the middle of main span from 1997 to 2005 are then analyzed to obtain the mean wind speed, the mean wind direction, the mean wind inclination, the turbulence intensity, the integral scale, and the wind spectrum of both 10-minutes and 1-hour in duration.
After identifying the wind environment surrounding the Tsing Ma Bridge, the next step is to analyze wind-induced bridge displacement response. Nevertheless, the GPS monitoring displacement data for the in-service Tsing Ma suspension bridge are induced by a combination of environmental and operational loadings, which include wind, temperature, and highway and railway traffic. In this connection, the bridge displacement response data recorded by the GPS during the period from 2002 to 2005 are collected, along with the temperature and vehicle flow data from the temperature sensors and weigh-in-motion sensors, respectively. Several algorithms are developed using MATLAB as a platform to pre-process the GPS measurement data for producing high quality databases. An identification method is subsequently developed to extract wind-induced displacement response by eliminating temperature-and traffic-(highway and railway) induced displacements and GPS background noise from the measured total displacement response. The relationship between wind speed and wind-induced displacement response in the lateral and vertical directions is finally explored according to wind direction and the locations of GPS stations on the bridge deck. The results show that the magnitude of displacement response varies with wind direction, and the location of GPS stations. The relationship between wind speed and displacement response is almost quadratic in the lateral and vertical directions. However, the aforementioned relationships are limited to the locations where GPS receivers are installed. In addition, the maximum wind speed encountered by the bridge and measured so far is also smaller than the design wind speed. Therefore, how to perform serviceability and strength assessments of the bridge is a challenging issue. This necessitates the integration of computer simulation of the bridge under the action of wind and WASHMS-based measurements. For this purpose, a complex structural health monitoring-based finite element model (FEM) for the Tsing Ma Bridge with significant bridge deck modeling features is used. The wind forces, composed of steady-state wind loads due to mean wind, buffeting forces due to turbulent wind, and self-excited forces due to interaction between wind and bridge motion, are then generated and distributed over the bridge deck surface following a series of procedures. The displacement responses of the bridge are then computed and compared with the responses measured from the field. The comparison is found to be satisfactory in general. The statistical relationship predicted from the field at the mid-main span is thus extended to extreme wind speeds and other locations on the bridge deck through computer simulation. The results are finally compared with measurements data from wind tunnel tests and the allowable movements of the bridge under the given limit state for serviceability assessment. The outcomes demonstrate that the lateral and vertical displacements follow the designed pattern and the bridge functions properly when the bridge is subjected to strong winds. As the span length of a cable-supported bridge increases, the bridge may suffer considerable buffeting-induced vibration across a wide range of wind speeds over almost the whole design life of the bridge. The frequent occurrence of buffeting responses of a relatively large amplitude may result in serious fatigue damage to steel structural components and connections, and may subsequently lead to catastrophic failure. Therefore, in addition to assessing its serviceability, it is also imperative that the strength of a long-span cable-supported bridge under high winds be assessed on the basis of wind-induced stress/strain analysis. However, the number of sensors available to take strain measurements on the Tsing Ma Bridge is limited. It is not possible to monitor all the stress responses of all the local components directly. In this connection, the complex structural health monitoring-based FEM which replicates the geometric details of the as-built complicated bridge deck, is used to compute wind-induced stress in all the bridge components and identify critical steel members. The wind-induced stresses of the critical members are then linked to the wind-induced displacement response at the mid-main span through the hybrid use of the GPS measured displacements and FEM analyses. The wind-induced stresses derived at extreme wind speeds are then compared with the yield stress of the steel material to assess the strength of the bridge. The results demonstrate that the mean stresses, stress standard deviations, and total stress responses vary quadratically with mean displacement, displacement standard deviation, and total displacement, respectively. The outcomes also demonstrate that the strength of the Tsing Ma Bridge under strong winds is guaranteed.
Subjects: Hong Kong Polytechnic University -- Dissertations
Bridges -- Aerodynamics
Bridges, Long-span -- Mathematical models
Bridges, Cable stayed -- Mathematical models
Global Positioning System -- Mathematical models
Pages: 1 v. (various pagings) : ill. (chiefly col.) ; 31 cm.
Appears in Collections:Thesis

Show full item record

Page views

44
Last Week
0
Last month
Citations as of Mar 24, 2024

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.