Combine Successive Sources Deletion Based on Kurtosis and Energy with ANC for Trackside Acoustic Weak Axle Bearing Signal Enhancement

The trackside acoustic system plays a crucial role in monitoring and diagnosing the status of axle bearing. However, due to the strong impact interference in trackside environment, the diagnosis accuracy has declined. The paper presents a new successive sources deletion approach based on the kurtosis and energy estimators (SSD-KE), which achieves the weak signal of axle bearing enhancement even in case of strong rolling noise. Additionally, a combination of SSD-KE and adaptive noise cancellation (ANC) is considered since the signal from axle bearing itself contains the component of rolling noise, which caused by the small distance and the rigid connection structure between axle and wheel-rail contact. The application cases of combined method are illustrated, which involves the validations of effectiveness and the anti-noise property by comparing with existing method and simulating two signal-to-noise ratios (SNRs), respectively. Moreover, the filtering parameter is also studied to investigate its effect on diagnosis result.