Application of wavelet exponential window denoising and dynamic uncertainty in acoustic emission

Abstract

The gas-liquid two-phase acoustic emission (AE) signal contains rich flow information, but it is also accompanied by a large number of interference signals. To accurately extract the characteristics of gas-liquid two-phase flow, the removal of interference signals is very important. In this paper, AE technology is used to detect the signal of gas-liquid two-phase flow in a vertical pipeline. The support degree of the sensor is checked by the trust function to confirm the consistency of the sensor and eliminate wrong data. The decomposition level of the wavelet base and wavelet transform is determined by four parameters such as the signal-to-noise ratio. By comparing the wavelet exponential window smoothing method and the wavelet soft threshold method, the wavelet exponential window smoothing method which is more suitable for the denoising effect is selected, and the real-time denoising effect is evaluated by using the measurement dynamic uncertainty theory. The results show that the wavelet exponential window denoising method improves the signal-to-noise ratio, reduces the energy leakage during denoising, and significantly improves the pseudo-Gibbs phenomenon, while dynamic uncertainty can effectively evaluate the denoising effect of AE signals.