Construction of optimal dual-tree complex wavelet for bonding quality detection of explosive composite structure

摘要

This paper investigates the construction method of optimal dual-tree complex wavelet (ODTCW), to extract the desired mode feature signal from the vibration signal without mode aliasing. Dual-tree complex wavelet with attractive properties such as shift invariance and reduced spectral aliasing may provide a possibility to extract the desired features from vibration signal. However, the fixed basis functions independent of the analyzed signal may weaken the advantage of the method and even decrease the accuracy of detection result. To overcome this deficiency, the proposed construction method starts with the design of approximate hilbert transform pairs of wavelet bases, by means of changing the parameters of vanishing moment K and the degree of approximation to the half-sample delay L, a set of wavelet basis functions with different frequency characteristic obtained. Based on the analyzed signal, by the rule of the crest maximum principle to search for optimal dual-tree complex wavelet basis functions which best match the analyzed signal. The ODTCW is combined with time entropy to detect bonding quality of explosive composite pipes and the results demonstrate its superiority.