Damage Detection of Structures with Detrended Fluctuation and Detrended Cross-Correlation Analyses
Abstract »Recently, fractal analysis has shown its potential for damage detection assessment in engineering field such as biomedical and mechanical engineerings. Due to its viable capability for interpreting irregular, complex, and disordered phenomenon, a structural health monitoring (SHM) system based on fractal analysis, namely Detrended Fluctuation Analysis (DFA) and Detrended Cross-Correlation Analysis (DCCA) to detect damage condition and location was proposed. Firstly, damage condition can be swiftly detected by analyzing ambient vibration signals measured from structure through DFA. Secondly, by analyzing the cross-correlation signals of different floor via DCCA, damage location can be determined. Moreover, a damage index was proposed based on multi-scale DCCA curve to improve the accuracy of damage location. In order to verify the performance of the proposed SHM system, a four-story numerical model was used to simulate with different noise effects. Furthermore, experimental verification was also carried out on a seven-story benchmark structure at National Center for Earthquake Engineering (NCREE) to assess the potential damage. The result showed that DFA method can detect the damage condition confidently, and the damage location can be identified by analyzing the cross-correlation signals of different floors through DCCA analysis with an accuracy rate of 75%. Moreover, based on the result of the damage index method, damage location can be correctly assessed with an average accuracy rate of 87.5%. The proposed SHM system is a promising application in practical implementation.