Research on a sound-based method for belt conveyor longitudinal tear detection

Aiming at the problems of accuracy, non-real-time performance, poor reliability and high complexity in belt conveyor longitudinal tearing detection, a belt conveyor longitudinal tearing detection method based on sound is proposed. The sound signals of belt conveyors are collected using a sound acquisition processor. First, the sound signal is processed by a preprocessing algorithm. The characteristics of the processed sound signal are extracted using the LFCC and GFCC combined with a feature extraction algorithm. The feature data of the sound signal are divided into a training set and a test set. The parameters of the convolutional neural network are improved by training and optimization of the training set data. Finally, a convolutional neural network model of the sound signal characteristics of a belt conveyor is obtained. The model is used to classify and identify the longitudinal tearing sound signals of the test set. The experimental results show that this method can realize the identification and detection of the longitudinal tearing sound of a belt conveyor. The average accuracy of the detection is above 94%, and the average processing time is 21.6 ms. The detection range is large, the cost is low, and the accuracy, real-time performance and reliability are high, which meets the requirements of longitudinal tear detection. © 2022