Date of Award

1988

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Civil and Environmental Engineering

First Advisor

George Z. Voyiadjis

Abstract

A structural identification methodology propounded by the author is used to identify the modal parameters and synthesize the flexibility matrix of a region of a prototype concrete bridge. Rigorous analytical investigation is employed first. This helps determine critical parameters of the structural model and guide the test procedure so that the best possible data can be attained. The bridge region is then tested using low-level, easily induced impact excitation. This is found to produce broad-banded response with a relatively high signal-to-noise ratio. Methods are suggested to optimize the test procedure so that it can be more systematically applied and so that more accurate data can be acquired from it. Modal parameters, except damping ratios, are first identified then the mass matrix is estimated and the flexibility matrix is derived. Model parameters are altered rationally and a very good match between experimental and model flexibilities is attained. This is done despite the existence of several closely spaced modes and the use of a simple identification algorithm (FFMI, Hogue, 1987). Modifications to the algorithm are explored and many are found useful in improving its performance. A static test is conducted to validate the dynamic test-based identification. Several problems and sources of error are associated with this test but a very strong agreement is shown between static and dynamic test results. An identification based on moving traffic excitation is also attempted and is shown to be generally inaccurate and unreliable. Analytical modeling, field testing, and data reduction are found to be highly interrelated tasks. Further research must be undertaken to generalize the integration of them. In particular, general test section- non-test section continuity modeling principles must be established. Improvements to the FFMI algorithm, particularly in estimating structure mass, or implementation of some other algorithm must also be explored in order to develop a more automated identification procedure. In addition, the optimal number of measurement locations must be determined. Furthermore, ways to conduct identification at different stress levels must be developed. Finally, the applicability to other structure types of the identification procedure used must be examined.

Pages

411

Share

COinS