The development of a systematic experimental method for damage identification

Abstract

The diagnostics of slight damage are extremely significant for providing the early warning damage information for in-service structures. This thesis presents the development of a systematic experimental method to identify structural damage by the experimental techniques. Three carbon fabric squared composite p1ates were used as the research objects. Two of them with light crack damage that can be classified as fiber breaking and matrix cracking were supposed to be identified through the dynamic experimental techniques. The tests of the frequency response functions (FRFs) of the investigated objects were conducted first to provide a general understanding of the dynamic properties of the material and the structures. Then the tests of the velocity fields at some specified frequencies are performed to acquire dynamic response data of the objects for the study purposes.

A systematic method to process the experimental data has been developed first in this thesis. The best regressive mathematical models for the test velocity fields are built based on the linear polynomial regression procedures and statistical analysis. To perform damage identification, the correlation coefficient (CC) and spatial correlation coefficient (SCC) techniques based on the best-fitted models and the curvature models were used. Finally, the student t' statistical tests were applied to decide whether the two compared data sets are significantly different in statistical sense.