Extracting procedures of key data from a structural maintenance database

Abstract

Maintenance continues for structure’s life cycle, which usually costs a lot. Inspection and/or monitoring are widely implemented to investigate the conditions of the structures. Existing databases are sometimes referred to for understanding the performance of structure with inspection/monitoring data. Since structural performance is related to various irregularly time-shifted factors, it is complicated to analyze a database efficiently. To improve the situation, the key data selection (KDS) method was proposed to extract key data from a database in this paper. The KDS method is first introduced in the context of conventional maintenance procedure. Afterwards, the detailed principle, implementation procedure, and possible applications of the KDS method are explained. Verification is performed with the data from the natural vibration of a bridge girder produced by vehicles (normal conditions) and from the lateral displacement of a bridge girder during a typhoon (accidental conditions). The results showed that the method can lessen the amount of data without changing their tendency and keep the error within an acceptable range. As a result, the KDS method is simple, reliable, and capable of reducing data under normal and accidental conditions.