Non-Destructive Detection of Wire Rope Discontinuities from Residual Magnetic Field Images Using the Hilbert-Huang Transform and Compressed Sensing.

Juwei Zhang,Xiaojiang Tan,Pengbo Zheng

doi:10.3390/s17030608

Juwei Zhang, Xiaojiang Tan + Show 1 more

Open Access

https://doi.org/10.3390/s17030608

Copy DOI

Abstract

Electromagnetic methods are commonly employed to detect wire rope discontinuities. However, determining the residual strength of wire rope based on the quantitative recognition of discontinuities remains problematic. We have designed a prototype device based on the residual magnetic field (RMF) of ferromagnetic materials, which overcomes the disadvantages associated with in-service inspections, such as large volume, inconvenient operation, low precision, and poor portability by providing a relatively small and lightweight device with improved detection precision. A novel filtering system consisting of the Hilbert-Huang transform and compressed sensing wavelet filtering is presented. Digital image processing was applied to achieve the localization and segmentation of defect RMF images. The statistical texture and invariant moment characteristics of the defect images were extracted as the input of a radial basis function neural network. Experimental results show that the RMF device can detect defects in various types of wire rope and prolong the service life of test equipment by reducing the friction between the detection device and the wire rope by accommodating a high lift-off distance.

Highlights

Wire ropes are widely employed components in diverse areas, such as in industrial production, tourist cable cars, mining, metallurgy, shipbuilding, and elevators
residual magnetic field (RMF) images of various wire ropes were acquired using the prototype device, and the above discussed procedures were applied to detect the occurrence and location of wire defects in addition to the number of broken wires included within the defect
In comparison with previously published results using only wavelet filtering in conjunction with the use of the shape characteristics of the defect RMF image and a back propagation (BP) neural network [22], the novel filtering algorithm proposed in this paper is more adaptive because the Hilbert-Huang transform (HHT) is applicable to any unknown signal, while wavelet filtering suffers from the substantial disadvantage regarding an appropriate choice for the wavelet basis

Summary

Introduction

Wire ropes are widely employed components in diverse areas, such as in industrial production, tourist cable cars, mining, metallurgy, shipbuilding, and elevators. Wire rope is a heavily loaded component, and long-term continuous operation eventually result in corrosion, wear, broken wires, loose wires, and fatigue, which decrease the loading strength of the rope, and can cause accidents, resulting in property damage and injury [1]. The development of a fast, non-destructive, and automatic detection technology is necessary. Wire rope defects include three main types: the loss of metallic area (LMA), local faults (FLs), and structural faults (SFs). The main non-destructive testing (NDT) methods employed for wire rope inspection include electromagnetic detection, ultrasonic guided wave (UGW) evaluation, radiation testing, eddy current inspection, and optical detection [1].

Methods

Results

Conclusion