53 / 2025-03-29 20:24:50

Research on identification method of conveyor abnormal condition based on Convolutional neural network

broken chain fault of stuck chain; Permanent magnet synchronous motor drive; Current change characteristics; Loss function optimization

Aiming at the difficulty of online detection of the broken chain of scraper conveyors, this paper presents a method to identify the broken chain of scraper conveyors based on permanent magnet synchronous motor-driven current change characteristics. Through the rigid-body dynamics-discrete element and electromechanical control co-simulation method, the rigid-electro-mechanical coupling model of scraper conveyor is constructed, and the transient sudden change characteristics of motor current signal in the conveyor running state and stuck chain state are studied. The motor current signal under the operating condition of the scraper conveyor was collected on-site. Wavelet transform and convolutional neural network analysis methods were used to extract and identify the motor current signal characteristics under the operating condition with a broken chain. The elastic network was used to optimize the convolutional neural network loss function, and the accuracy of the identification model of the broken chain was verified.