Data-Driven Damage Detection for Beam-Like Structures under Moving Loads Using Quasi-Static Responses

摘要

Due to safety problems of modern structures continuously becoming more serious, structural health monitoring (SHM) systems are attracting increasing attention. The huge amounts of data being generated by SHM systems every day are challenging to analyze. Due to data-driven methods only considering the changes in signals, they are well-suited for this problem. This study considered an application of a data-driven method, moving principal component analysis (MPCA), for damage detection and location for moving-load quasi-static responses of a beam. The MPCA is an adaption version of principal component analysis (PCA) in which a fixed-length window is used to limit the data length so that it has some advantages in adapting to damage detection. Both finite element model and experimentation were used to prepare the time series for detection. It was noted that the MPCA algorithm showed good performances on both numerical and experimental data.