Doppler Correction Using Short-Time MUSIC and Angle Interpolation Resampling for Wayside Acoustic Defective Bearing Diagnosis

Abstract

Wayside acoustic defective bearing detector (ADBD) system plays an important role in monitoring the condition of high-speed train bearings and ensuring their normal operation. However, as the system acquires acoustic signals from a passing train by a fixed microphone array, the acquired signals are severely distorted by the Doppler effect, which is an obstacle to defective bearings detection. This paper proposes a novel scheme by combining a short-time multiple signal classification (MUSIC) and angle interpolation resampling (AIR) to remove the Doppler distortion embedded in the microphone array signals. Two basic steps are included in the proposed scheme. First, the proposed short-time MUSIC method is performed on the array signals to calculate out the time-varying receiving angle information. Then, the proposed AIR method is conducted to correct the distorted signal. Compared with the traditional schemes, the proposed scheme is simple and hardly needs prior knowledge, which is expected to be applied in the ADBD system. The proposed scheme is verified by means of simulation and practical studies. Results indicate that it has superior performance to remove the Doppler distortion and has obvious advantages for the ADBD system.