A DEMON Line Spectrum Detection Method Based on Parameters Pre-estimation

摘要

In order to enhance the ability of extracting characteristics from short time propeller radiated-noise signal, a detection of envelope modulation on noise (DEMON) method based on parameters pre-estimation (PE-DEMON) is proposed according to the model of multiple composite hypothesis testing. In the PE-DEMON method, the feedback mechanism between the data preprocessing and shaft frequency estimation is designed using the harmonic characteristics of the DEMON line spectrum. In the process of PE-DEMON, several preprocessing parameters are supposed and the spectrum analysis is performed, then pre-detection of the line spectrum and pre-estimation of the shaft frequency is made based on the spectrum data. The parameters of data preprocessing are modified by the output of the pre-estimator. The line spectrum detection and the shaft frequency estimation is carried out again to obtain the final output. The simulated data analysis show that the PE-DEMON method can improve the estimation accuracy of the shaft frequency and the detection probability of line spectrum compared with the traditional DEMON method.