Rapid nondestructive assessment of the degrees of sensitization of 5456 aluminum alloys using laser-induced breakdown spectroscopy (LIBS) with multivariate analysis
Langue
en
Article de revue
Ce document a été publié dans
Journal of Analytical Atomic Spectrometry. 2023, vol. 38, n° 7, p. 1494-1500
Royal Society of Chemistry
Résumé en anglais
The 5xxx series aluminum (Al)–magnesium (Mg) alloys are extensively used in marine transportation and naval ships due to their high strength-to-weight ratios and great corrosion resistance properties. However, material ...Lire la suite >
The 5xxx series aluminum (Al)–magnesium (Mg) alloys are extensively used in marine transportation and naval ships due to their high strength-to-weight ratios and great corrosion resistance properties. However, material degradation of the 5xxx alloys occurs over time after exposure to environments at elevated temperatures, referred to as sensitization. Sensitization is characterized by β-phase (Mg2Al3) precipitates along the grain boundaries, which increase the susceptibility to intergranular corrosion (IGC) and stress corrosion cracking (SCC). Therefore, it is important to assess the degree of sensitization (DoS) in modern naval ships, whose structures are primarily made of Al–Mg alloys, to determine whether they have reached a DoS threshold that may require maintenance or repair activities. In this study, laser-induced breakdown spectroscopy (LIBS) with statistical multivariate analysis methods was successfully developed to assess the DoS levels of the 5xxx alloys. LIBS measurements were conducted after chemical etching on 5456 Al–Mg alloy samples with DoS values of 7.1 (unsensitized), 20.2 (indeterminate), and 47.3 (sensitized) mg cm−2, respectively. The statistical multivariate analysis method of principal component and discriminant function analysis (PC-DFA) was used to process the LIBS spectra to establish a calibration model used for prediction of Al alloys with unknown DoS values. For each alloy sample, a total of 400 LIBS spectra were acquired. 350 of them were used to establish the calibration model by PC-DFA and the remaining 50 were used as unknown spectra for external validation of the model established. External validation of the established classification model shows accuracies of 100%, 100%, and 82%, respectively, for predicting tested Al alloy samples with unsensitized, indeterminate, and sensitized DoS values. Successful external validation of the classification model established by PC-DFA indicates the ability for sensitization assessment of 5456 Al alloys with unknown DoS values. The technology developed has the capability for rapid and non-destructive sensitization assessment on site.< Réduire
Origine
Importé de halUnités de recherche