Research On Immune Danger Theory Based Fault Diagnosis Method Of Hydraulic Pump

The hydraulic pump has been compared to the heart of the hydraulic system byhydraulic engineer,its health will directly affect the normal work of the hydraulicmanufacturing equipment.Hydraulic pump failure,it may affect some functions of thehydraulic manufacturing equipment,It also maybe cause serious,catastrophic accidents inproduction.Therefore to research state monitoring and fault diagnosis technology on thehydraulic pump is particularly important.At present,the fault diagnosis technology alongthe direction of the development of automation,intelligent.Inspired by the biologicalimmune danger theory,for reducing dimension from a complex information of theoriginal high dimensional feature vector,this paper proposed feature selection algorithmbased on Danger Immune Theory.This paper proposed the fault diagnosis algorithmbased on immune danger theory principle which characteristics of development withlearning,clustering,memory characteristics of,and this algorithm is applied to faultdiagnosis of hydraulic pump.At present,there are not many scholars at home and abroad apply Immune DangerTheory to fault diagnosis field,but this theory is applied in information security,machinelearning,data mining and other fields.This paper based on recognition mechanism ofDanger Immune Theory of biological,apply Matlab software to develop a high dimensiondata dimension reduction algorithm and fault diagnosis algorithm.Feature selectionalgorithm is reduced to a low dimensional vector form the complicated information ofhigh dimensional vector,greatly reducing the follow-up time of fault diagnosis.Faultdiagnosis algorithm will form the memory antibody population by learning the sample asantigens,Fault diagnosis is memory antibody population to test samples.In order to verify the effectiveness of the algorithm,based on the laboratory materialtest machine with axial plunger hydraulic pump as a diagnostic object.Apply theacceleration sensor and NI data acquisition card acquire vibration signals of hydraulicpump,Analysis original vibration signals of hydraulic pump and determine t the resonantfrequency range by zoom spectrum analysis technology.according to the resonant period of signal,using wavelet cluster based envelope demodulation methods for determining envelope demodulation; the envelope signal is demodulated2layer wavelet decomposition and reconstruction, extracted from each subband signal reconstruction in the time domain, frequency domain and time domain and frequency domain information of the original feature vector; the choice of objective function (the between class scatter matrix samples the trace and sample within-class scatter matrix trace ratio) as a feature selection feature subset classification performance evaluation function, application based on Danger Immune Theory feature selection algorithm, the objective function value of the maximum corresponding to the feature vector; at last,based on Danger Immune Theory fault diagnosis based on after the feature selection of the learning samples (antigen) for learning,and generate the final of each state of mature detector (memory antibody population) in order to complete the test samples (antigen) of state monitoring and fault diagnosis.Through the Matlab software simulation,verification effectiveness based on Danger Immune Theory Algorithm for fault diagnosis of hydraulic pump.