Research On Intelligent Fault Diagnosis Method Of Double-tooth Roller Crusher Based On Multi-parameter

The paper is launched as an important sub-project of the key coal-based science and technology research project “Self-moving crushing station and key technology development(No:MJ2014-02)”.The project started in September2015.The National and Local Joint Engineering Laboratory for Mine Intelligent Electrical Appliance Technology of Taiyuan University of Technology is one of the cooperation units.This is the main task to effectively monitor and diagnose the double-toothed roller crusher in self-moving crushing station,and timely grasp the machinery.The normal operating status ensures the process of coal mining production,and improves the safety,reliability,efficiency,energy saving and environmental protection.In view of the fact that the semi-continuous mining process of China’s self-propelled crushing plant has just started,the theoretical and experimental research is not yet mature.At the same time,with reference to the harsh mine geological environment in open pit mining in China,combined with the existing fault diagnosis technology,the double-toothed roller crusher is studied.This paper,from the mechanical structure and working principle,analyzed the common fault types of its key parts,collected the data information that can reflect the state of the crusher,designed the monitoring and fault diagnosis platform based on WinCC,The platform would lay the foundation for the digitization and intelligence of large-scale open pit mining.In particular,the time-frequency image texture feature analysis technology is introduced to study the fault diagnosis algorithm of the rolling bearing of the double-tooth roller crusher.The specific work content is as follows:By consulting the literature and on-site investigation of the geological environment of the semi-continuous mining process of large open-pit coal mines,we can understand the working mode of the double-toothed roller crusher in the self-moving crushing station.Then the paper analyzed the structure,working principle and fault situation of the double-toothed roller crusher.The analysis results,especially for the transmission system,can thus determine the running states and fault types of the double-toothed roller crusher.Introduced the image texture feature extraction technology,this paper analyzed the fault operation and mechanism of the rolling bearing,and proposed a bearing fault feature extraction method based on time-frequency image.Firstly,the EEMD＿WVD algorithm combining with EEMD and Wigner-Ville time-frequency analysis was used to obtain the time-frequency three-dimensional image for the rolling bearing vibration signal with uncross-interference and high time-frequency.Then,the circular LBP was used to enhance the image texture features.At the same time,the time-frequency LBP spectrum was obtained,and the gray histogram was extracted as the feature quantity.Finally,It can reduce the feature quantity using PCA.The low-dimensional feature quantity that can be used for the rolling bearing fault diagnosis.The 33005 tapered roller bearing in the 2PGC double-tooth roller crusher drive system was selected,and the bearing fault test device was built with relevant electrical equipment to collect the four vibration signals.Aiming at the weak fault pulse of bearing vibration signal,the time-frequency image feature extraction method based on EEMD＿WVD,LBP and PCA proposed in this paper was used to complete a series of operations:time-frequency characterization,feature extraction and data dimension reduction.Dimensional feature quantities would be input BP neural network to realize bearing fault classification,complete fault identification of four bearing states of crusher.The fault diagnosis accuracy can be as high as 99%,which verifies the effectiveness of bearing fault diagnosis method proposed in this paper.After clarifying the overall requirements and design ideas of the system developed by the project,the target functions of the system design include data acquisition,data communication,data analysis,and status monitoring&diagnosis.According to the determined system monitoring quantity,the appropriate sensor should be selected to realize the collection of crusher voltage,current,temperature,pressure,liquid level,vibration and other informations,then S7-300 PLC and PCI analog-to-digital conversion was used to convert the collected analog quantity.It is digital and uploaded to the server via Industrial Ethernet communication.With Siemens WinCC as the mainstay and SQL Sever & MATLAB as the supplement,this paper had designed a friendly human-machine interface for crusher condition monitoring and fault diagnosis.The interface functions mainly included condition monitoring,fault diagnosis,parameter setting,alarm record,and history record.Using OPC communication technology,the interface between WinCC and MATLAB was realized by instruction form.The system fault diagnosis function can be perfected by MATLAB,which owned the powerful data processing capability.The built-in crusher condition monitoring and fault diagnosis platform was used to realize processing analysis and storage memory for the crusher data.The monitoring and diagnosis of the crusher was completed.