Method On Off-line Fault Diagnosis And State Evaluation Of Urban Rail Vehicle Axle Box Bearing

As the key component of urban rail vehicle, performances of axle box bearing will directly affect the vehicle’s safe operation. With the speed increasing of urban rail transit, the increasing of passenger flows, and the operating conditions of bogie are more severe. Besides the immature development of on-line measurement, low accuracy, high cost, great maintenance work, all which bring safety concerns to the axle box bearing. so to carry out research on the axle box bearing off-line vibration fault diagnosis and state evaluation, it’s can improve the accuracy of the fault diagnosis, and has important meaning to avoid major mistake and change the maintenance strategy.In the face of the characteristics of the axle box bearing on-line detection and difficulty online axle box bearing fault feature extraction, and according actual needs, this paper studies related theory and technology which in view of the axle box bearing fault diagnosis and state evaluation. The main work is:(1) The analysis of structure characteristics for urban rail vehicle axle box bearing and the vibration mechanism has been finished. Through analyses axle box bearing to the structure characteristics, common failure types and causes, the vibration mechanism and failure characteristic frequency, etc. which provides the theory basis for the method of urban rail vehicle bogie axle box bearing diagnosis.(2) the analysis of stress of the two states for the urban rail vehicle axle box bearing static and run and has been finished. The dynamical model of single degree of freedom fault vibration impact and two degree of freedom vibration impact for Axle box bearing. Based on structure characteristics, failure mode, and dynamic model for axle box bearing, we analysis the signal fault vibration characteristics. It provides a theoretical support for the study of the axle box bearing fault vibration diagnosis method.(3) The scheme of the axle box bearing off-line detection has made. According to characteristics of axle box bearing vibration, the primary testing solutions has verified on-site, through analysis comparison to sensor acquisition signal of the axle box bearing different position, finally the optimal sensor placement scheme is determined.(4) Research and established a set of based on improved Empirical Mode decomposition(Ensemble Empirical Mode Decomposition, EEMD)-Hilbert envelope demodulation method of axle box bearing fault diagnosis. By analyzing the advantages and disadvantages of the methods in the existing time domain, frequency domain and time-frequency analysis for the aspect of bearing test, finally determined the method which is based on improved EEMD- Hilbert envelope demodulation method. On the basis of traditional Hilbert envelope demodulation, introducing the EEMD decomposition. Through setting the decomposition parameters and analysis the Intrinsic Mode Function component on the double correlation coefficient analysis and calculating the threshold value with the correlation coefficient, it improves the efficiency of the IMF resolution and analysis. Finally, it verifies the accuracy of the algorithm with of axle box bearing failure data.(5) The method which is based on morphological spectrum of dimensionless parameter axle box bearing condition assessment is proposed. It is verified that recognition ability of morphological spectrum to the type of signal time domain waveform with the simulation. And the shortage of the traditional time domain dimensionless parameter has been analyzed, ultimately determines the dimensionless parameter state evaluation method based on pattern spectrum. Using the real car test data validates, it has been verified that accuracy of the algorithm which State evaluation based on pattern spectrum dimensionless parameters.