Electromagnetic tomography for defect detection of carbon fiber cable using Krylov subspace

Abstract Carbon fiber cables are gradually used as large-span structural cables in bridges, amusement rides and other fields, and the on-line detection technology of the damage during its service is becoming more and more important. Electromagnetic tomography (EMT) is used for non-contact detection of surface defects on carbon fiber cable. In this paper, a novel image reconstruction algorithm based on the Krylov subspace is used to improve the imaging effect of the surface defects of carbon fiber cable. The image reconstruction process of EMT is transformed into the constrained least squares problem, and twofold regularization is realized for the original problem. The projection data is maximized and the optimal solution is searched in the constructed space. Simulation results have shown that the Krylov method outperforms the other methods with better indicators such as image error and correlation coefficient and has better anti-noise performance. The experimental studies are carried out by the EMT system with a double layer sensor array, which has shown that the method has better imaging effect.