Wire Rope Damage Detection Based on Magnetic Leakage and Visible Light

Quankun Chen,Juwei Zhang,Qiang Ye
DOI: https://doi.org/10.1007/s11668-023-01682-z
2023-05-27
Journal of Failure Analysis and Prevention
Abstract:This paper designs a set of wire rope broken wire damage detection systems, which realizes the complete process of wire rope damage identification, location, cutting, and classification. The system uses the designed two-dimensional magnetic flux leakage (MFL) detection device and line array camera to collect the MFL signal and visible light signal of the steel wire rope, respectively, and reduce the noise of the MFL signal through the filtering algorithm combined with modified ensemble empirical mode decomposition (MEEMD) and wavelets. The magnetic signals in two dimensions are fused by a guided image filtering method, which realizes the enhancement of MFL images. Then, a wavelet transform-based convolutional autoencoder (WTCAE) is proposed for visible light defect detection and localization. Finally, a convolutional neural network is used to identify broken filaments. The experimental results show that the system has a high detection sensitivity.
What problem does this paper attempt to address?