WAVELET-BASED HEALTH MONITORING OF RANDOMLY EXCITED STRUCTURES

Abstract eng:
This paper concerns the wavelet approach for structural health monitoring in which the wavelet transform is used as a detector for the singularities in the structural responses, which are caused by sudden changes in the system parameters. If the system is subject to broadband random excitation like earthquakes, however, the singularities could be caused by the excitation process, too. Therefore, some isolation method is needed to separate the singularities due to the excitation and the singularities due to the system’s changes. First, the input-output correlation is investigated in the wavelet domain in order to explain how the singularities in the input propagate through the system and affect the output. The derived equation shows that the wavelet transform of the output at a small scale can be decomposed into a superposition of the wavelet transforms of the input and the components corresponding to the system’s changes. A simple numerical example shows, using this equation, the influence of the input’s randomness is successfully removed and the spike indicating the time of the system’s change is extracted.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket