Status Analysis And Fault Diagnosis Of The Rotating Equipment Based On Vibration Testing

Most of the mechanical equipment is rotating machinery; it covers the electric powerindustry, chemical industry, metallurgy, machinery manufacturing and other important fields.The running speed of rotating machinery is higher, and often are the key equipment of thefactory, such as generator, turbine, blower, heavy sizing-press and so on, the working state notonly affects the safe and stable operation of the machine, but also directly effect the follow-upproduction. Some of the fault occurred in the early stages, vibration quantity was small, whichwas often drowning in other types of the vibration. If you can’t find problems and timely takemeasures to solve the problems, it will lead to serious consequences. Therefore, these largerotating machines are the main research objects on condition monitoring and fault diagnosis,and the equipment state monitoring is particularly important.Rotating machine is the machine that depends on the rotor movement, and it must havethe most basic of rotor, bearing parts and so on on the structure. The rotor is one of the mostimportant compoents of the rotating machinery, so it has very important significance to rotor’scondition monitoring and fault diagnosis. This paper is based on vibration test, and hasidentifying analysis and fault diagnosis on the SJ14500spindle blower rotor of Baotou Steeliron three burn workshop. This paper used a variety of signal processing methods included thetime region analysis, frequency region analysis, wavelet and wavelet packet time-frequencyanalysis and neural network diagnosis method. From the analysis of time domain, based onthe time domain features statistics, combined with neural network, established the time domainBP network and had state recognition on blower rotor; from the analysis of frequency domain,expounded several spectrum analysis methods, and based on the spectrum energy of spectrumanalysis, combined with the network, established the frequency domain BP network, and hadstate recognition on blower rotor; in the time-frequency domain analysis, bases on the band’srelative energy which after the wavelet packet decomposition and reconstruction, combinedwith the neural nerwork, established the wavelet BP network, and had state recognition onblower rotor. Through the effective combination of the neural network with three traditionalsignals processing methods, baseing on the analysis and processing of a large number ofmeasured data, verified the feasibility of the network model to identify the blower rotor state.