Impact of parameters on damage indicators of structures: Symbolic data analysis and principal components

Abstract

A number of techniques have been proposed for the detection of damage to structures. Some of these techniques are based on vibrational data, whereby changes in physical properties can be identified through alterations in vibrational characteristics. The majority of techniques are based on variations of modal parameters. The direct utilization of accelerations offers considerable advantages in damage detection, making the process more rapid and more straightforward by eliminating the necessity for an identification stage. The paucity of research into raw dynamic data (accelerations) can be attributed to the considerable volume of data that must be stored and manipulated. Some authors suggest Symbolic Data Analysis (SDA) and Principal Components Analysis (PCA) as techniques for reducing the volume of data without compromising its fundamental characteristics. The objective of this study is to examine the number of categories utilized in SDA and the number of principal components in PCA that yielded the optimal damage state classification results when applied to experimental data from a simply supported beam and numerical data from a four-story Benchmark. Two techniques for determining the number of SDA categories are put forth for consideration.