Find in Library

Stress corrosion cracking (SCC) widely occurs in the key structures of large-scale mechanical systems and may cause catastrophic accidents without evident macroscopic deformation. A sufficiently accurate quantitative evaluation of SCC is important to guarantee the safety and efficient operation of large mechanical systems. However, the reconstruction accuracy of SCC shape parameters using eddy current testing (ECT) signals remains a problem, given the complex microstructure of cracks and the properties of the ECT inverse problem. In this study, a two-dimensional numerical model of crack with conductivity distribution is used to calculate the ECT signals due to SCC. The relation between crack parameters and ECT signals is investigated through simulation and experiment. Results show that the ill-posedness of SCC shape reconstruction is mainly manifested in the nonuniqueness of the solution of crack shape parameters. Therefore, a regularization method based on the features of multifrequency ECT signals is proposed to solve ill-posedness by adding prior information to the solution space of the optimization problem. The reconstruction results of the simulated and measured ECT signals from the conductive crack models verify that the proposed strategy can effectively improve the precision of SCC shape reconstruction.