Time Synchronous Averaging Based on Cross-power Spectrum

Abstract Periodic signals produced by rotating machinery is of great significance for its fault diagnosis and health monitoring. Time synchronous averaging is an effective and convenient technique for extracting those components. However, the performance of time synchronous averaging is seriously limited when the separate segments are poorly synchronized. In order to address this problem, a new averaging method that is capable of extracting periodic components without external reference and accurate period is proposed in this paper. With this approach, the phase differences of all segments can be basically eliminated by phase detection and compensation. Therefore, the segments to be averaged have a high synchronization. The efficiency of this method is validated by numerical simulation and experimental investigation. The proposed method can improve the adaptability of time synchronous averaging.