Early Fault Detection of Gearbox Using TSA and VAR Model Considering Load Variation

Early fault detection and prognostics and health management (PHM) are of importance in the gear transmission system. Most of the existing fault detection methods rarely considered the influence of load variation. A novel gear fault detection scheme based on time synchronous averaging (TSA) and vector autoregressive (VAR) is proposed, which enables to implement early fault detection of a gearbox under varying load. Six time series models are established for different load conditions. The gear residual signal is selected to reduce the interference of variable conditions to real signals. The root mean square, variance and kurtosis are used to analyze the degradation trend of the gearbox. The results show that the proposed method can detect the abnormal condition of the gearbox 35 flies ahead of the traditional method, which verifies the effectiveness of the proposed approach.