Damage identification of steel-ECC composite deck incorporating piezoelectric impedance methodology and hierarchical clustering analysis.

As a highly ductile concrete, engineered cementitious composites (ECC) can be used as pavement to form a lightweight composite bridge deck system. However, the structural damage introduced by fatigue load in operation might lead to the degradation of structural performance. In this paper, piezoelectric sensors and hierarchical clustering algorithm are used to identify structural damage of steel-ECC composite deck. First, three steel-ECC composite decks were tested under four-point loading, and the electrical impedance signals were measured. The root mean square deviation (RMSD) was extracted to quantify the structural damage severities and locations. Then the frequency interval is divided into nine sub-frequency range to employ the sensitivity analysis. On this basis, a hierarchical clustering algorithm was introduced to analyze the impedance signal to identify the damage of steel-ECC composite deck. The results show that the development of the structural damage can be continuously monitored using impedance methodology and hierarchical clustering algorithm even in the case of small unlabeled datasets. [ABSTRACT FROM AUTHOR]