Research On Fault Feature Extraction Method Of Reciprocating Compressor Valve Based On ITD And Maximum Entropy Spectrum Analysis

As the core mechanical equipment in the petrochemical industry,reciprocating compressors,if they fail,bring huge economic losses to the enterprise,and may also lead to casualties and serious accidents,ensuring safe and stable operation is the focus of equipment maintenance.Therefore,research on fault diagnosis of reciprocating compressors has great practical significance and economic value.In recent years,domestic and foreign scholars have made great progress in the research of fault diagnosis of reciprocating compressors.For the problem that the fault characteristics of reciprocating compressors are difficult to extract under strong noise background,this paper takes 2D12 reciprocating compressor as the research object,and targets reciprocating compressors.The fault diagnosis technology has carried out a series of research work.Firstly,on the basis of reviewing the development history of fault diagnosis technology for reciprocating compressors,the common technical methods for fault diagnosis of reciprocating compressors and the problems of each method are described.The adaptive method for non-stationary signal decomposition is expounded.Secondly,the end effect,over-decomposition and spurious component appearing in the simulation analysis and the measured signal of the reciprocating compressor are analyzed,and the corresponding causes are proposed.The specific performance is:(1)Abandoning the traditional image continuation method for processing endpoint effects,a method of bidirectional extension and setting of separation regions at both ends of the signal is proposed.Compared with the traditional method,the effect of suppressing the endpoint effect is better.(2)The termination condition of the ITD decomposition is that the residual signal is a monotonic function.In order to satisfy this condition,there is often a problem that excessive decomposition leads to a large number of false components.Aiming at this phenomenon,it is proposed to use the cumulative variance contribution rate as the basis for the termination of ITD decomposition,to improve the operation speed and reduce the generation of false components;(3)For the phenomenon that the ITD decomposition algorithm generates false components and the non-stationarity of the reciprocating compressor test signal,a method combining correlation entropy and piecewise cross-correlation analysis is proposed to remove the false component,which can retain the true component to the maximum extent.False weight;(4)There are high-frequency noise and other interference factors in the test signal of the reciprocating compressor,which leads to the distortion of the PRC component after the decomposition of ITD.A signal preprocessing method for improving the noise reduction by the translation-invariant multi-wavelet adjacent coefficient method is proposed.Finally,according to the advantages of the maximum entropy spectrum analysis for short data analysis and the characteristics of the valve fault vibration signal,the fault feature extraction method of the reciprocating compressor valve based on ITD and maximum entropy spectrum analysis is proposed.Experiments have shown that due to the different state of the valve,the characteristic frequency of the extracted gas valve is also significantly different.