Fault Diagnosis of Pipeline Blockage in Residual Membrane Recovery Machine Based on VMD-ReliefF-BP Neural Network

During operation, the integrated residual film and cotton stalk recycling machine often suffers from blockages in the conveying channel due to excessive feeding, high humidity of cotton stalks, and excessively long cotton stalk fragments, which significantly reduces the efficiency of agricultural production. Aiming at the problem of frequent blockages in the cotton stalk conveying channel of the integrated residual film and cotton stalk recycling machine during cotton stalk recycling, this paper proposes a diagnostic method based on variational mode decomposition, feature dimension reduction, and BP neural network. This method employs Variational Mode Decomposition (VMD) to decompose the signal Then, the main IMF components are selected based on the correlation coefficient criterion and the signal is reconstructed. subsequently, features in the time domain and frequency domain are extracted, and feature signal selection is conducted using the Relieff algorithm. Finally, the selected features are used as inputs for the BP neural network to achieve identification of blockage faults. The simulation experimental results demonstrate that the fault diagnosis rate of the BP approach is 95%, indicating its effectiveness in diagnosing faults and laying the foundation for further research.