Fault Diagnosis Method Of Chuck Head Of Chemical Fiber Winder Based On Deep Learning

In the winding and forming process of chemical fiber production,the winder is responsible for winding and forming the post-processed virgin fibers,where the winder chuck,as a direct winding component,has a direct impact on the quality and productivity of chemical fiber cakes.As chemical fiber equipment becomes more and more complex and sophisticated,the traditional method of relying on manual experience for troubleshooting does not meet the needs of intelligent production.Therefore,it is important to study the intelligent fault diagnosis method of winding machine to improve the reliability of card head winding and ensure the quality of chemical fiber products.However,due to the strong background noise interference and the variable speed phenomenon in the winding process of the chuck,the fault diagnosis of the winding machine chuck becomes a difficult engineering problem to solve.In this paper,to address these difficulties,we propose a winder chuck fault diagnosis method based on signal processing and deep learning,taking winder chuck as the research object.The main research work is as follows:Aiming at the problem that the chuck head vibration signal contains strong noise interference,resulting in weak fault characteristics,this paper proposes a fault diagnosis method for fixed speed chuck head bearings based on EMD envelope spectrum and convolutional neural network.Through the signal adaptive decomposition and reconstruction criterion,the components containing the fault signal are screened out to achieve signal noise reduction processing;a chuck head bearing fault diagnosis model is built,and the envelope spectrum of the reconstructed signal is used as the input of the model to identify the fault type through fault feature extraction and classification.The validation diagnosis accuracy on three different data sets is higher than 94%,which proves the effectiveness of the proposed method.Aiming at the problem of the variable speed of the chuck,this paper proposes a fault diagnosis method for the bearing of the variable speed chuck based on the order spectrum and dual-channel convolutional neural network.The order spectrum analysis is used to convert the frequency spectrum at different rotational speeds into an order spectrum that does not contain rotational speed information,which eliminates the influence of the fundamental frequency,so that a single diagnostic model can be used for the fault diagnosis task of chuck bearings at different rotational speeds;using dual-channel convolutional neural network model extracts fusion fault features from different channels of the same signal acquisition point.In the experimental verification stage,the data collected at three rotational speeds were used to train the model,and other rotational speed data were used for verification,the fault diagnosis accuracy was higher than 98%.In the comparison test phase,the diagnostic accuracy of the proposed model in this paper improved by 3.1% over the conventional single-channel model.Taking the winder of a chemical fiber production plant in Zhejiang as the research background and integrating the above diagnostic models,a prototype system for fault diagnosis of the chuck bearing of the winder was developed.The prototype system has the functions of data acquisition,data processing,fault diagnosis,equipment status monitoring and so on.The research results of this paper provide theoretical methods and technical tools for the fault diagnosis of chemical fiber winder chucks.The example verification shows that the prototype system can effectively improve the accuracy and efficiency of the fault diagnosis of the winder chuck,and improve the intelligence level of chemical fiber equipment.