A comprehensive condition monitoring system for steel hoisting wire ropes based on machine learning and synchronized signal processing of the optical and magnetic channels
Abstract
A comprehensive condition monitoring system for steel hoisting wire ropes based on machine learning and synchronized signal processing of the optical and magnetic channels
Incoming article date: 05.01.2026A comprehensive approach is proposed for automated diagnostics and condition monitoring of steel hoisting wire ropes, implemented by integrating two independent methods—optical and magnetometric—into a single synchronized monitoring system. In the optical channel, two analysis mechanisms are implemented: defect classification based on evaluating characteristic patterns of changes in the cross-sectional dimensions, and classification using a convolutional neural network trained on annotated images of real damage. The magnetometric channel applies the magnetic flux leakage principle, detecting internal anomalies using a sensor array whose signals are converted into a numerical feature vector. Temporal and spatial synchronization of the data using correlation algorithms provides unified defect mapping and minimizes false alarms. Experimental validation was conducted on ropes with defects such as bending, kinking, and breakage, as well as on undamaged ropes, under conditions close to real operation. The results confirm high sensitivity, noise robustness, and the potential suitability of the proposed solution for continuous industrial monitoring.
Keywords: automation of monitoring, steel wire ropes, non-destructive testing, integrated monitoring, computer vision, defect classification, neural networks, gradient boosting, convolutional neural networks