Fault Diagnosis Method Of Reciprocating Compressor Based On Optimal Quality Factor RSSD And HFE

Reciprocating compressors are widely used in petroleum and chemical industries.They are mainly responsible for the compression and transportation of dangerous gases such as coal gas,natural gas,and ethylene.Once internal failures occur,they will not only cause major economic losses,but also pose serious threats to personal safety.Therefore,the fault diagnosis and detection of reciprocating compressors has become a hot topic.The structure of the reciprocating compressor is complex and there are many internal excitation sources.The vibration signal exhibits strong non-stationary,nonlinear,and multi-component coupling characteristics.In addition,the working environment of the reciprocating compressor is complex,and the vibration signal is often accompanied by noise signals.The vibration state information is submerged in these complex vibration signal,how to filter out the useful fault information from the composite signal is the focus of the fault diagnosis of the reciprocating compressor,and it is also a difficult problem that has been researched for a long time.In view of the deficiencies in fault extraction of reciprocating compressors,this paper proposes a fault diagnosis method based on the combination of optimal quality signal RSSD and HFE.Different from traditional frequency division signal decomposition methods,the RSSD method constructs different wavelet basis functions according to different quality factors,and then uses morphological analysis methods to establish the sparse decomposition objective function,and then obtains corresponding coefficients through iterative methods.It can realize the effective separation of signals with similar center frequencies and overlapping frequency bands,forming high and low resonance components.The method of this paper is applied to the fault diagnosis of reciprocating compressors,and the faults are identified and classified by SVM.The results show that this method can accurately express the fault information and effectively diagnose the failure of the reciprocating compressor.First,consult the literature,understand the structure of the reciprocating compressor,the working principle and the common failure mechanism,describe the development history and research status of the reciprocating compressor fault diagnosis,and understand the algorithm flow and implementation steps of the signal RSSD.Secondly,through the in-depth study of signal RSSD theory,a hierarchical hybrid optimization algorithm combining genetic algorithm and particle swarm optimization algorithm is proposed for the problem of traditional manual selection of quality factor signal decomposition problem.The algorithm adopts a hierarchical structure,the bottom layer uses agenetic algorithm,and contributes to the global search capability;the top layer uses a particle swarm algorithm to accelerate the convergence speed.The disturbance vibration signal of the reciprocating compressor is mixed with a large amount of disturbance information,and the layered hybrid optimization method can achieve effective separation of the disturbance components fault impact and noise by adaptively selecting the quality factor.Simulation experiments and measured data show that the optimal quality factor signal RSSD method can effectively analyze non-stationary signals.Again,introduce the basic theory and algorithm steps of HFE.HFE can measure the complexity of signals at different nodes through hierarchical and coarse grain analysis.The algorithm fully considers the influence of coarse graining on spectrum division,and effectively divides the time series according to spectrum characteristics.It can not only analyze the low-frequency components of the signal but also analyze the high-frequency components of the signal,thus avoiding information omission due to improper construction.The frequency spectrum of the vibration signal of the reciprocating compressor is complex,and the characteristics of the signal cannot be accurately described from a single angle.HFE calculates the fuzzy entropy of the hierarchical sequence obtained by each node by constructing different frequency bands of the signal,and then the entropy values of different segments of the same node are optimized to accurately and comprehensively describe the signal characteristics.Reciprocating compressor fault experimental data shows the superiority of HFE in feature extraction.Finally,take the 2D12 type reciprocating compressor as an example to simulate the failure of the bearing and the valve respectively,and use the fault diagnosis method of the best quality factor signal RSSD and HFE presented in this paper to diagnose and identify,the method first uses the hierarchical structure to optimize the quality factor,and performs the signal RSSD,calculates the low-resonance component HFE,constructs the feature vector,and then uses the support vector machine for classification and identification.The results show that the proposed method has high recognition accuracy and can effectively diagnose different faults.