Research On Condition Monitoring And Fault Diagnosis Method Of The Rotary Kiln Based On Vibration Analysis Of Supporting Rollers

Rotary kiln is a typical large slow-speed running mechanical equipment.As key equipment,rotary kiln is widely used in the cement and metallurgical industry.In China,there are more than 1,000 cement rotary kilns.The problems such as kiln crank,overload of the supporting rollers,hot bearings,cracks in the supporting rollers seriously influence the safety operation of the rotary kiln.For a long period,the break maintenance and periodic maintenance are the mainly maintenance strategy for the rotary kiln,and it lacks the real-time condition monitoring system.Moreover,the monitoring theory of the rotary kiln has not been studied.In order to reduce the maintenance cost,minimize the amount of downtime and lengthening the service life,it is urgent to develop a real-time condition monitoring and fault diagnosis system for the rotary kiln.Therefore,the research on condition monitoring and fault diagnosis method of the rotary kiln based on vibration of supporting rollers was conducted in this thesis,and the main researches are listed as follows:A rotary kiln cylinder deformation detection method is proposed as it is the main reason causing the vibration of the supporting rollers.Firstly,the characteristics and causes of cylinder deformation are analyzed.Secondly,according to the deformation characteristics,a cylinder deformation and three-dimensional mode model calculation method are proposed.Finally,experiments in the real industry field are conducted,and the experimental results indicate that the proposed measurement method is effective,which providing the studying base for the vibration mechanism of the supporting rollers.Considering that the vibration model of the support rollers can be used to provide theoretical guidance for extracting the effective information and fault pattern recognition of the kiln,a vibration model of the supporting rollers under the internal and external incentives was analyzed,and the model was verified by the experiments in the industry field.Moreover,the response characteristics of the supporting rollers under the different fault modes of the rotary kiln were analyzed.The result indicated that the running state of the kiln can be effectively reflected by the vibration monitoring of the supporting rollers.In order to extract the fault characteristic information of in the vibration signals,a novel strategy was proposed based on WTD-CEEMD with parameter optimization method.Firstly,an optimization parameters method of CEEMD method was proposed.Secondly,the signals were denoised by the WTD method,and the CEEMD with parameter optimization is employed to decompose the signals and the sensitive IMF components were selected.Finally,the signals were constructed based on the sensitive Intrinsic Mode Functions(IMF)components,the operation states of the kiln can be evaluated by the energy of the constructed signal.The proposed method is verified with the designed measurement system in the actual industry field.And the experiment results indicated that the proposed method is effective to identify the operating condition rotary kiln.With the increase of the data collected by the sensors during the process of monitoring,it is becoming difficult to identify the vibration modes of the supporting roller of the rotary kiln using the tradition signal processing method.It is urgent to develop a reliable and efficient fault diagnosis method.And an intelligent fault diagnosis approach for supporting rollers using multiclass SVM with parameter optimization and PCA for feature reduction was proposed in this paper.Firstly,the fault related time-frequency features were extracted by the CEEMD method as that the vibration signals of the supporting rollers usually contain background noise,which significantly affects the results of fault diagnosis.Also,another twenty time domain and frequency domain were extracted.Secondly,the extracted feature was multidimensional reduced through the principal component analysis(PCA).Finally,in order to find the optimal parameters for the multiclass SVM model,the particle swarm optimization(PSO)was used.The proposed method was verified with the experiment data collected by the monitoring system in field experiment.The result indicated that the method could effectively identify the vibration mode of the supporting rollers.Based on the above research,a real-time monitoring and fault diagnosis system of rotary kiln was developed and designed.Firstly,the rotary kiln proactive monitoring theory was introduced,the options of monitoring parameters as well as the measuring points and sensors were analyzed,and a monitoring system was developed.The system which is able to conduct data acquisition,signal analysis and fault diagnosis.Finally,the designed system was verified by the experiments.Meanwhile,in order to meet the short-term and distributed data acquisition for kiln monitoring,a data acquisition system based on wireless sensor networks(WSN)was developed,and the system was tested in both the laboratory and the industry field.