Localization of Acoustic Emission Source Based on Improvement LMSTDE

Abstract

A number of strong noises usually cover up the useful acoustic emission (AE) signals when the units is rubbing between the rotors and surrounding stationary components, and the attenuation is serious during the signals transmitting in metal. Such phenomena make it difficult to achieve the location of AE source exactly. In order to solve this problem, the least mean square time delay estimator (LMSTDE) is introduced to estimate the distance between two sensors. The new algorithm adds an adaptive gain estimator, both time delay and gain are refreshed by the bias simultaneously. In order to speed up the convergence, the gradient of error is taken to adjust the variable step size in the case of large deviation between estimated signals and reference signals, On the contrary, by adopting forgetting factor, the deviation is taken to adjust the next step to decrease the Mean Square Error (MSE). The experiment results show that the new algorithm is prior to the traditional generalized cross correlation (GCC) time delay estimate in location accuracy and what is more the algorithm has a better performance in convergence by applying the variable step size.