Fault Diagnosis Of The Coiling Tension Control Process

The coiling tension control is to guarantee constancy of the coiling tension without influence of increasing and decreasing speed of the mill in the coiling process.Too much or too little tension will have an impact on the final product quality,particularly on the flatness when the coiling machine winds strips.Therefore,the fault diagnosis in coiling tension control system is not only available to the maintenance,but also crucial to the guarantee of the product quality.This paper sets the coiling tension system as its research subject,adopting the fault diagnosis method based on data driven to conduct diagnosis in the system.The main research contents are as follows:1)Explaining the research background and significance,this paper gives an introduction to the definition and classification of faults,general procedures of faults diagnosis and classification of diagnostic methods.In combination with the need of the equipment itself,this paper also puts forward the fault diagnosis technique based on the data driven to conduct diagnosis over the system.2)This paper introduces the constitution of this system,analyzing three methods to the coiling tension control,namely the indirect control method,the direct control method and the composite control method.In allusion to the indirect control method,the paper makes an analysis of factors affecting the coiling tension control.3)The paper also introduces the fault diagnosis technique based on data driven.This technique is to perform analysis and process over processing data so as to finish the system fault diagnosis without any awareness of the accurate analytical model.This sort of methods can be divided into machinelearning,statistic analysis,signal processing and so on,while the partial least squares(PLS)is one kind of the statistic analysis method.This paper puts emphasis on giving an introduction to the principle,arithmetic,feature distraction,test index and other aspects of the PLS.4)The paper gives an introduction to the general process of the fault diagnosis through application of the PLS.What's more,this paper collects signal data having an influence upon the coiling tension control process,conducting detection and diagnosis over the three common faults in the process with the PLS.The simulation result indicates that even though this diagnosis method has some false positives,on the whole it is of accuracy and feasibility.5)Finally,the paper introduces the theory of wavelet transform.The paper makes an analysis of the theory and implementation method of the wavelet de-noising technology,and applies the technology into the fault diagnosis in the process.The simulation result shows that the processed signal data are much easier to detect and diagnose faults than the unprocessed ones.