Fault Diagnosis Of Cement Rotary Kiln Based On Multivariate Data Drive

Cement plays an important role in our modern economic construction,and is an irreplaceable raw material for capital construction,industrial construction and other related fields.In the cement clinker calcining system,the rotary kiln is one of the most core equipment.Its importance is self-evident.As the production process structure is more and more complex and the working environment is harsh and the thermal parameters are numerous,the real-time industrial data in the cement production process shows complex data characteristics,and the fault diagnosis is more and more difficult to rely on manual realization.How to solve the complex structure of process data while improving the operating efficiency of enterprises has become a problem worth deep thinking.In order to ensure the safe production of cement rotary kiln and improve the operation efficiency of enterprises,this thesis takes cement rotary kiln as the research topic,and investigates and studies its process flow,process parameters,process data and management mode.Aiming at the multi-modal characteristics of rotary kiln process data,a new method based on Local Information Local Neighbor Standardization and Principal Component Analysis(LLNS-PCA)is proposed to establish a high precision fault diagnosis model.Firstly,the process data is decomposed into several local samples by Gaussian mixture model method,and the mean value and variance of each local sample are normalized by nearest neighbor.Then,principal component analysis is used for fault monitoring.Based on PCA monitoring model,T~2and SPE monitoring statistics were used to monitor multi-modal processes.Finally,the penicillin production process and actual production data of cement rotary kiln were used to verify its effectiveness.In this thesis,a process monitoring information management system for cement rotary kiln is designed by combining the technological flow of cement rotary kiln with the existing automation level of enterprises,using Django,Vue and My SQL database.Through eight functions of user management,data management,model training,model management,workshop monitoring,device monitoring,team rating and daily report,the system completes the collection,analysis and management of equipment operation data,realizes the accurate fault diagnosis of rotary kiln,and solves the pain points of factory production operation and maintenance.Finally,the functional test and non-functional test of the system are carried out,and the system meets the user’s requirements and achieves the expected results.The research of this thesis provides a feasible solution of real-time condition monitoring and fault diagnosis system for the safe operation of cement rotary kiln.This scheme can provide a certain reference for the control and optimization of cement enterprises in the production process.At the same time,it also provides a meaningful reference value for the chemical industry to promote the intelligent application of production and manufacturing industry.