Status Monitoring and Fault Diagnosis of Cigarette Machine Feeding System Method Study

In order to achieve real-time monitoring of the status of the cigarette machine feeding system and automatic diagnosis of faults, a method for monitoring and diagnosing the status of the cigarette machine feeding system based on multi-sensor application and machine learning is proposed. The time-frequency characteristics of the suction pipe negative pressure data are extracted through wavelet packet analysis, and the average and standard deviation of the suction pipe negative pressure data are used as feature vector inputs. The performance of convolutional neural network and support vector machine in fault classification and recognition is compared. The experimental results show that the state recognition method based on support vector machine (linear kernel) can effectively diagnose faults in the cigarette machine feeding system, laying a theoretical foundation and technical support for the online monitoring and fault diagnosis of the cigarette machine feeding system.