Research On Early Fault Diagnosis Of One-way Valve Of Reciprocating High Pressure Diaphragm Pump

Reciprocating high-pressure diaphragm pump is the core power equipment of a long distance,high-lift,high-concentration slurry transportation pipeline,whether its working status is normal or not directly affects the production efficiency of enterprises.Due to the influence of high working frequency and complex mechanism,the check valve becomes the most vulnerable part in the high pressure diaphragm pump.The check valve is easily affected by the particle size distribution of the conveyed minerals and the rheological properties of the slurry,which leads the failure of check valve to show the characteristics of sudden,concurrency,multi source,uncertainty and so on,at the same time the vibration signals of check valve also show non-stationary and nonlinear characteristics.In the early failure of a check valve,the impact component that reflects the characteristics of the fault is very weak in the whole vibration signal,and it is easily affected by the influence of other components and environmental noise,which makes it difficult to detect and diagnose the early failure of the check valve.In view of this problem,the following studies have been carried out in this paper:(1)Aiming at the problem that the traditional time domain decomposition method can't separate the similar frequency components and the weak high-frequency components,an early fault detection method of check valve based on Directional Empirical Mode Decomposition(DEMD)and K-L divergence is proposed.The vibration signal of fault using DEMD algorithm for noise reduction and decomposition,and select the effective component signal through the K-L divergence valuel,using the Hilbert marginal spectrum of instantaneous spectrum analysis of the fault feature information of the signal,the experimental analysis shows that this method can preliminarily extract the early fault characteristics of the check valve,which is feasible and effective;(2)Aiming at the problem of reducing the noise and reducing the useful signal by using the DEMD method,an the early fault detection method of check valve based on Stochastic Resonance(SR)and DEMD is proposed.Variable scale pretreatment onthe fault vibration signal,and input it to the SR system optimization after the optimized parameters,the output signal of the DEMD decomposition algorithm,using the envelope spectrum analysis of fault information,experimental analysis shows that this method can avoid the disadvantages of the traditional decomposition method,can extract early fault features better;(3)Based in feature extraction methods,in order to improve the classification results of the single valve fault types,a fault diagnosis method based on kernel principal component analysis(Kernel Principal Component Analysis,KPCA)and least squares support vector machine(Least Squares Support Vector Machines,LSSVM)is proposed.KPCA is used to extract the two time feature of the time and frequency domain parameters,fault classification in the LSSVM fault identification model which is input to the optimized parameters.The experimental results show that this method can improve the recognition accuracy of the check valve.In this paper,the check valve is taken as the research object.Aiming at the problem that the fault characteristics of the check valve are not obvious and the fault feature information is difficult to extract,a series of targeted methods are used to extract and classify the early fault features of the check valve.