The existence of false components with the Hilbert vibration decomposition (HVD) method has seriously restricted its application in practical rotor fault diagnosis. To solve this problem, an improved HVD method was proposed by adopting Kullback-Leibler (K-L) divergence values as a distinguishing index of true and false components, which is named the KL-HVD method. First, it calculated the K-L divergence values between the HVD components and the original signal, and then, these values are compared with the set threshold. Finally, it eliminated the false components whose K-L divergence values were larger than the threshold. The experimental results of rotor fault signal analysis demonstrated that the KL-HVD method could more accurately extract the timefrequency characteristics of the faults and the K-L divergence value was more suitable as the distinguishing index of true and false components than the mutual information and correlation coefficient values.
In order to get the compressibility factorZof working fluid under different conditions, experimental measurement method ofZunder high pressure and high temperature and data mining method were studied in this paper. Experimental measurement method based on real gas state equation and prediction method based on Least Squares Support Vector Machine were proposed. First, an experimental method for measuringZat high temperature and high pressure was designed; in this method the temperature, pressure, and density (mass and volume) of corresponding state were measured and substituted into the actual gas equation of state, and thenZcan be calculated. Meanwhile, in order to obtain continuous value inT-pplane, Squares Support Vector Machines are introduced to establish the prediction model ofZ. Take Hexamethyldisiloxane, for example; the experimental data ofZwas obtained using the experimental method. Meanwhile the prediction model ofZ, which can be used as calculation function ofZ, was established based on those experimental data, and theZ(T: 500 K~800 K,p: 1.3 MPa~2.25 MPa) was calculated by using this calculation function. By comparison with this published data, it was found that the average relative error was 2.14%.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.