Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/71584
Title: GPS-based multi-sensor systems for structural health monitoring
Authors: Yang, Wentao
Advisors: Ding, Xiaoli (LSGI)
Keywords: Structural health monitoring
Sensor networks -- Design and construction
Issue Date: 2017
Publisher: The Hong Kong Polytechnic University
Abstract: Structural health monitoring (SHM) is to use sensors to extract structural characteristics information, such as stress, strain, deformation, displacement, velocity and acceleration. However, the deformation signals are usually very weak and contaminated by noise, as they distribute in different space and time domains, and are acquired by different ways and different sensors. Therefore, it is necessary to process them to get the characteristic information which is sensitive to structural damage. Based on GPS observations, this thesis brings in other sensors and establishes proper mathematical models to form an integrating algorithm for structural health monitoring, which can combine the information obtained by more than two sensors to achieve a higher observation accuracy. The thesis first analyzes the problems in GPS, tilt-meters, and accelerometers when applied to monitoring structural health. It then proposes some integration methods for combined use of these sensors. The main innovations of the thesis are as follows, 1) A platform for integrating the multi-GPS antenna and tilt-meter instruments has been studied. The priori information that the distance between instruments is a constant, is taken as constraints and added to the observation equations of GPS and the tilt-meter, in which a new data processing model is proposed. Verification experiments show that when the GPS observation is abnormal, the integration algorithm can improve the precision by 37%, and the improvement in the elevation direction is the most significant.
2) This study proposed to combine stacked multi tilt-meters and GPS for applications such as monitoring high-rise buildings and landslides. A new least squares algorithms with constraints are proposed. An adaptive factor is given to adjust the contribution of the GPS observation and the tilt-meter observation to the parameter estimation, which provides a new way for multi sensor data fusion. The case study show, when the GPS observation is abnormal, the standard deviation of results obtained by the adaptive fusion algorithm was reduced by about 40% in all the three directions (north, east and elevation). 3) Based on an analysis of existing methods for fusing GPS, accelerometer and tilt-meter observations in monitoring structural dynamics, a Kalman filter-based model with constraints is established for integrating the different sensors data. Experiment results show that the new algorithm can improve the precision by 60% when the GPS observation is abnormal and accuracy is poor (less than 10mm). 4) An adaptive filtering algorithm with constraints is presented for the fusion of GPS, accelerometer and tilt-meter when the weighting scheme is iteratively estimated. Simulation experiments with different GPS observation precisions and different numbers of satellites in simulation experiments show that the proposed algorithm can significantly improve the reliability of the monitoring system, especially when the GPS observation is abnormal and satellites are insufficient. The thesis offers some new techniques for real-time monitoring of structures such as dams, high-rise buildings, and slopes for protecting lives and properties.
Description: xx, 178 pages : color illustrations
PolyU Library Call No.: [THS] LG51 .H577P LSGI 2017 YangW
URI: http://hdl.handle.net/10397/71584
Rights: All rights reserved.
Appears in Collections:Thesis

Files in This Item:
File Description SizeFormat 
991022020058203411_link.htmFor PolyU Users167 BHTMLView/Open
991022020058203411_pira.pdfFor All Users (Non-printable)4.26 MBAdobe PDFView/Open
Show full item record

Page view(s)

21
Citations as of Feb 19, 2018

Download(s)

4
Citations as of Feb 19, 2018

Google ScholarTM

Check


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