Experimental study on classification and identification method for fault signals of mining steel wire ropes

Abstract

In the coal mining production process, wire ropes play a crucial role as the main tool for transporting personnel and materials in the entire coal mine hoisting system, and their fault identification is particularly critical. To address the problem of fault identification of mining wire ropes, a recognition model design combining K Singular Value Decomposition (SVD) with Genetic Algorithm (GA) optimized Support Vector Machine (SVM) is proposed. Firstly, K Singular Value Decomposition is used to denoise the collected wire rope damage signals. Feature vectors are extracted from the denoised and reconstructed signals to form the feature set of wire rope damage. Genetic Algorithm is utilized to select the optimal parameters of the Support Vector Machine and train it to obtain the optimal identification model for wire rope damage. Simulation and experimental results show that compared with wavelet thresholding and Empirical Mode Decomposition (EMD) denoising algorithms, the method using K Singular Value Decomposition for denoising achieves a signal-to-noise ratio of 25.166dB and a root mean square error of 0.805. Compared with Particle Swarm Optimization (PSO) optimized Support Vector Machine identification model, the identification rate of the Support Vector Machine identification model optimized by Genetic Algorithm is higher, reaching 96.67%. This method realizes accurate identification of different damage signals of wire ropes, which is conducive to ensuring the safe and efficient production of coal mines.