Rolling Intelligent Fault Diagnosis Method Research And Application

With the development of modern science and technology, modern equipments are featured by more and more complicated structures, more and more perfect functions, more and more automaticity and larger system integration scale. As a result, there are more and more requirements concerning defect diagnosis and performance identification. As one of the most commonly used general spare parts, the rolling element bearings are also damageable ones. Therefore, people attach much importance to theory, methods and applications related to fault diagnosis of rolling element bearings.Some existing rolling element bearing fault diagnosis methods, which are mainly based on the basic principles and methods of vibration analysis, are analyzed and discussed in detail. The analysis methods of vibration signals are discussed thoroughly, including time domain analysis, frequency domain analysis and time-frequency analysis, as well as general approaches of extracting features from the time domain and frequency domain are introduced. Besides, two techniques of status identification, neural networks and support vector machine, are introduced. New methods of vibration signal analysis and automatic recognition of bearings' state are presented. The new methods are studied in depth theoretically, based on which codes are developed and then applied to bearings diagnosis with real data.The research work of this paper includes two parts:(1)Combined adaptive short-time Fourier transform and the principle of traditional resonance demodulation technique, a new method of vibration signal analysis—adaptive resonance demodulation method is presented. Time-frequency spectrum of the vibration signals collected is obtained through adaptive STFT, and time-energy signals are extracted according to L~p norm rule from it instead of the envelope of the filtered signals in traditional resonance demodulation technique. The application of the new method shows that it not only performs well in extracting features, but also avoids the difficulty of...