A PCA-fused multi-feature method for estimating the tension of steel strands based on pulsed eddy current testing

As a critical load-bearing component in many infrastructures, the tension status of steel strand serves as a key indicator for infrastructure safety assessment. This study focuses on estimating the tension of steel strands and proposes a multi-feature fusion approach based on Pulsed Eddy Current (PEC) testing. Firstly, the mechanism of PEC-based tension estimating for steel strand is analysed in detail, and PEC experiments on steel strand under different tension are carried out. The response of the PEC signals to tension is consistent with the results of the theoretical analysis. Subsequently, five features are extracted from the PEC signals for estimating the tension of steel strand, and their characteristics and performance in tension estimation are analysed. Finally, based on these features, a Principal Component Analysis (PCA)-based method for multi-feature fusion is developed to improve the accuracy of tension estimation. Compared with the features without using fusion method, the performance in tension estimation of the significant principal component with fusion feature has been significantly enhanced. The Root Mean Square Error (RMSE) of the estimation results with multi-feature fusion is 0.73% of Full Scale (FS), and the maximum is 1.41% FS.