Research On Fan Vibration Condition Monitoring And Fault Diagnosis Based On Real-Time Characteristic Values And Data Mining

Fan is widely used in petroleum,chemical industry,electric power,metallurgy,etc.With the operation of equipment for a long time,the probability of the failure is rising greatly which could make it stop production,cause huge property losses and potential safety hazard.Therefore,fan equipment condition monitoring and fault diagnosis has extremely important significance.This paper introduced detailed information on the development situation of rotating machinery condition monitoring and fault diagnosis technology at home and abroad and founded that the current research situation is: there are mature methods and techniques on the vibration signal monitoring at home and abroad.But the frequency domain method of fault diagnosis is still basically off-line because of the large amount of calculation.The diagnosis analysis needs to be made manually and various application methods are professional which leads to not easy to complete real-time calculation and online analysis judgment and the lack of online,intelligent methods and techniques.And with reduction of price of the in situ monitoring equipment,the installation of sensors on the fan to monitoring status becomes possible.Fan equipment operation parameters database is established by collecting full cycle fan operating time domain characteristic values and combination of big data mining methods makes it possible to establish a fault diagnosis model based on time domain characteristic value.The advantage of doing so is that the fault diagnosis model will be real-time and on-line because the characteristic values are also real-time and on-line.At the same time,the accuracy of fault diagnosis will improve continually as the improvement of the database and data mining method.This paper mainly conduct the following work to solve the problem of lacking condition monitoring in CAP1400 containment recirculating cooling fan:Introduction is made about the basic concept of vibration,typical research results of rotating machinery vibration fault and the vibration signal feature extraction and analysis method.Typical characteristic values are selected to make the concrete analysis according to the international and domestic industry standard which built the foundation for fault diagnosis methods;Second,the fan vibration condition monitoring platform was designed and developed.The function of the vibration condition monitoring platform was also verified and perfected.Fan characteristic values was calculated and stored which established database of fan vibration characteristic value;Finally,data mining was made on characteristic values to make characteristic value sensitivity level distribution of different vibration fault which would determine the potential relationship between different characteristic parameters and predict the fan running state.On this basis,the fan vibration fault diagnosis model was developed based on real-time characteristic values and the model was used in physical fault.The diagnostic method of this paper is effective on the test bed frame which indicates that the diagnostic method based on data mining can effectively complete the online analysis of vibration fault and diagnose of different fault cause.It has the engineering application value and popularization value to realize the real-time monitoring and vibration intelligent remote diagnosis in fan,pumps and other rotating machinery which are widely used in machinery industry.