| The main ventilator of coal mine is the core equipment of mine ventilation system which is used to deliver fresh air to the mine and exhaust gas and dust. It is an important guarantee of safety to production enterprises that the main ventilator working in regular situation. Therefore, the condition monitoring and fault diagnosis of main ventilator is of great significance.Based on the research background of main ventilator, the technology of fault diagnosis and development status of fault diagnosis is summarized in this thesis. According to the structural characteristics of main ventilator, the thesis researches its failure mechanism and designs vibration signal acquisition system. Closely combine with some common failure mechanism of bearing and rotator failure of main ventilator, the thesis researches a fault diagnosis method based on the vibration signal.The traditional time-frequency analysis methods such as Windowed Fourier Transform, Wigner-Ville Distribution and Wavelet Transform, which applications and limits of main ventilator is analyzed and introduced. The traditional time-frequency methods are still the signal analysis method that based on Fourier Transform. In the strict sense, they are only suitable for analysis of stationary signal and lack of characteristic of self-adaptive.For defects of the traditional time-frequency methods in analyzing non-stationary signal, Hilbert-Huang Transform is applied to analyze vibration signal and fault feature extraction by using the method of Intrinsic Mode Function Energy Entropy. SVM is proposed to distinguish status of main ventilator according to the characteristics of fault signal of the main ventilator, also taken into account the number of samples and diagnostic accuracy. In addition, in order to improve the training speed and accuracy of classifier, the Genetic Algorithm is applied for SVM parameter optimization. Finally, the effectiveness of the algorithm is verified by example of main ventilator fault signal analysis. |
