Research on Fault Feature Extraction of Gearbox Bearing Based on Improved Genetic Algorithm

Under the condition of high-speed operation, the bearing fault will rapidly develop into a serious accident, which makes the fault diagnosis must be advanced. Therefore, it is very necessary to find a suitable method to suppress the interference of discrete frequency components and complex modulation phenomena in the frequency spectrum, so that the bearing fault characteristic components can be highlighted. In this paper, a fault feature extraction method of gearbox bearing based on improved GA (genetic algorithm) is proposed. For the fault feature extraction of rolling bearing, before amplitude demodulation analysis, the fault diagnosis result can be effectively improved by using GA to separate gear signal from bearing signal and further realize its fault feature extraction. Boltzmann strategy in SA (Simulated Annealing) algorithm is added to GA to form an improved GA, so as to improve the global search ability of GA. The research results show that compared with the traditional mathematical morphological spectrum analysis method, the improved mathematical morphological spectrum method proposed in this paper has better ability to distinguish vibration signals with different modulation frequencies, and is stable and time-consuming, so it is effective as a feature vector extraction method.