SYSTEM IDENTIFICATION METHODS FOR DAMAGE EVALUATION OF EXISTING STRUCTURES
Abstract
The objective of this study is to develop a practical approach for damage assessment of existing structures using structural response data as measured. Although the main emphasis is on the safety evaluation of structures following the occurrence of an earthquake, there is no restriction regarding the use of the approach for structures which have experienced other kinds of loads. In this dissertation, several methods are developed for the identification of inter-story hysteresis behavior of high-rise buildings and the detection of changes in their main natural frequency. Then apparent and meaningful features of the identified hysteresis behaviors are found. Finally, a suitable and practical damage indicator is obtained from and correlated to the extracted features. Because there exist uncertainties in results from using system identification techniques, the third part of this study is devoted to the search for techniques with which such uncertainties can be dealt. In this thesis, two new system identification methods for the estimation of natural frequency and hysteretic behavior of nonlinear structures are presented. Numerical examples of two full-size existing structures and two one-tenth scale frame structures are given to demonstrate the effectiveness of these new techniques. From the identified load-deflection relationships, it is concluded that there exists a "soft-to-stiff" type of behavior for reinforced concrete structures subjected to earthquake excitations. The particular property of structures under large-amplitude and repeated loads is used to introduce a new measure of structural damage. This new damage index is called "slope ratio" and defined as the ratio of structural stiffnesses under low-amplitude to that of initial unloading under large-amplitude loads. Results of experimental dynamic tests are used to evaluate the creditability of the proposed damage indicator. Finally, Dempster and Shafer's theory, which gives a rational inference procedure for problems with uncertainty, is used to form an algorithm for combining the results of this approach with those of other knowledge sources. This combination rule is shown to possess the effect of compensating for the individual deficiencies of knowledge sources and helps in decision-making regarding the damage state of structures.
Degree
Ph.D.
Subject Area
Civil engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.