Gearbox is the core component of transmission machinery. Due to its high working intensity and complex environment, faults often occur. In the process of equipment fault prediction and health management (PHM),is the object that needs to be focused on. The gear of the main transmission component is one of the components prone to failure. When a gear fault occurs, the collected vibration signal will contain FM and am signals of rotating frequency, meshing frequency and frequency doubling. How to extract the fault characteristic signal from the original signal is the key to fault diagnosis. Under the condition of constant speed, the fault signal appears periodically. Through resonance demodulation, wavelet analysis, EMD (empirical mode decomposition) and other methods, the fault signal can be effectively extracted. These methods are similar to band-pass filtering, which can separate the fault information from the signal, and the analysis effect is better for a certain speed.
However, under the condition of variable speed, the fault frequency of gear will change with the change of speed, which makes the frequency spectrum appear frequency fuzzy phenomenon. The original method can not extract the fault information. In practice, the working conditions of gearbox are complex and changeable, and the speed change often occurs. The single fault feature extraction method under constant speed can not meet the needs of daily PHM monitoring, so it is necessary to extract and diagnose the fault features in real time during the operation of equipment. Therefore, the study of feature extraction method under variable speed is of great significance to expand the scope of application of PHM, enrich feature extraction methods, and enhance the reliability and stability of equipment operation. It is necessary to strengthen the research on fault feature extraction of gearbox gear under variable speed.
Order tracking technology is a common method to deal with this kind of variable speed problem. Order tracking can transform non-stationary time domain signal into stationary angle domain signal by resampling, so as to eliminate the influence of speed change. Since its appearance, order tracking technology has attracted the attention of many scholars, and there are abundant researches in related fields. Rpotter et al. Proposed the concept of computational order tracking in 1989. Later, krfyfe, e dmunk et al. Analyzed and demonstrated the effect of computational order tracking, which further confirmed the accuracy and effectiveness of the method. Cheng Junsheng et al. Proposed a fault diagnosis method based on the combination of computational order tracking and LMD (local mean decomposition); Wang Kuang et al. Analyzed the vibration signal of planetary gearbox through order analysis technology, which reflected the advantage of order analysis technology for fault diagnosis of rotating machinery; Peng Fuqiang et al. Proposed an adaptive time-varying filter order tracking fault diagnosis method, which belongs to the order tracking technology without tachometer, and can better solve the problem of order signal interference. Now the commonly used order tracking technology can be divided into three categories: Hardware order tracking (hot), computational order tracking (COT) and tachometer free order tracking (TOT).