Diagnosing Belt Conveyor Idler Faults with STFT and CNN

Belt conveyors are widely used for transporting bulk commodities, and idlers are essential components prone to frequent failures. The timely and accurate fault diagnosis of idlers is critical for ensuring the safe and efficient operation of belt conveyors. This paper proposes a method for diagnosing idler faults by combining Short-Time Fourier Transform (STFT) and Convolutional Neural Networks (CNN). The STFT is used to convert one-dimensional vibration signals into time-frequency representations, which are then processed by the CNN to classify the operational state of idlers. The proposed method is tested on vibration signals collected under various working conditions, including normal operation, bearing damage, and cylinder skin fracture. The CNN model, trained and validated using MATLAB, achieves high diagnostic accuracy, demonstrating its effectiveness in identifying different fault types. This approach enhances the reliability of belt conveyor systems by enabling prompt detection and maintenance of idlers.