Structure analysis and denoising using Singular Spectrum Analysis: Application to acoustic emission signals from nuclear safety experiments
Résumé
We explore the abilities of the Singular Spectrum Analysis (SSA) to characterize and denoise discrete
acoustic emission signals. The method is first tested on simulated data for which different types and
levels of noise are considered. It is then applied on real data recorded from nuclear safety experiments.
The results show an excellent ability of the SSA to characterize the corrupted signal and to detect structural
changes, even for low signal-to-noise ratio. For denoising purposes, the quality of the results
depends mainly on the separability between the source signal to be estimated and the noise. However,
whatever the case, the main components of the source signal are clearly identified when the components
associated with the noise are removed.