Principal Component Analysis of Eddy Current Signals Obtained from Steam Generator Tubes by Bobbin Probe

The model-based interpretation tools for eddy current testing (ECT) signals have been developed by the novel combination of neural networks and finite element modeling for quantitative flaw characterization in steam generator tubes. The performance of inversion system strongly relies on the databases that had been used in the implementation of the specific system. It is also widely recognized that features play the most important role in the interpretation of ECT signals. In the present work, a database was constructed using synthetic ECT generated by the finite element models and principal component analysis (PCA) was adopted in order to optimize the feature set of eddy current signals. The features with PCA improved the performances of signal interpretations. The excellent performance obtained in the present work demonstrates the high potential of the developed inversion tools as a practical interpretation of eddy current signals. Steam generator tubes are pressure boundaries that separate the secondary unit from the primary one in nuclear power plants, so that they play a critical role in the safe operation of nuclear power plants.