Intelligent Fault Diagnosis of Broken Wires for Steel Wire Ropes Based on Generative Adversarial Nets

The quantitative identification of broken wires is great significance to maintain the safety mechanical systems, such as steel wire ropes. However, in order achieve high accuracy recognition results, a large number fault samples are necessary, which difficult practical industrial detection. In this paper, novel approach, based on generative adversarial nets (GANs) and convolutional neural network (CNN), proposed solve problem situations where real inspections have generated only small sample for analysis. One-dimensional original signals transformed into two-dimensional time-frequency images by continuous wavelet transform (CWT). Next, these used various defects combing GANs CNN with limited samples. main innovation paper that can be improved generating through GANs. experimental results demonstrate method achieves better rates compared existing detection methods.