Methodology based on statistical features and linear discriminant analysis for damage detection in a truss-type bridge

摘要

Corrosion is considered one of the main mechanisms that can damage a civil structure, especially truss structures. Hence, corrosion identification in its early stage is of vital importance. This paper proposes a vibration-based structural damage detection methodology which uses a statistical feature extraction, linear discriminant analysis, and a neural network classifier for assessing the healthy condition of the bridge, as well as its damage conditions produced by corrosion with three levels of severity. The obtained results show that the new method can identify the structure condition and the severity level due to corrosion with high accuracy, even when an incipient fault is present.