Abstract
The principles and features of the application of statistical principal component analysis (PCA) in active thermal testing are considered. A comparison between PCA and Fourier analysis in finding defects in composite materials, detecting corrosion in aluminum, and determining moisture content in construction materials is performed. It is concluded that, generally, images of principal components increase the signal-to-noise ratio and are close in performance to phase diagrams; nevertheless, the results of this method are poorly predictable and require further analysis.
Original language | English |
---|---|
Pages (from-to) | 509-516 |
Number of pages | 8 |
Journal | Russian Journal of Nondestructive Testing |
Volume | 44 |
Issue number | 7 |
DOIs | |
Publication status | Published - Jul 2008 |
ASJC Scopus subject areas
- Mechanical Engineering
- Mechanics of Materials
- Condensed Matter Physics
- Materials Science(all)