A Linear Narrowband Filtering Method Based on Migration-Enhanced Singular Value Decomposition and Its Application to Reduce Noise

With the development of the aviation industry, the limit of aeroengine noise is becoming increasingly stringent. Vibration mode extraction is the most commonly used method for reducing the noise of compressor, which is the main source of aeroengine noise. In vibration mode extraction, inherent errors occur due to the inherent defects in the measurement and analysis. To solve the problem, the acoustic pressure signal, which is among the most common and accurately measured signals and contains much useful information, is selected for analysis to eliminate the inherent errors. In addition, to compensate for the lack of analysis methods, based on the characteristics of the singular value decomposition (SVD) and acoustic pressure signals, a linear narrowband filtering method based on migration-enhanced singular value decomposition (LNF-MSVD) is proposed to accurately extract the acoustic mode from acoustic pressure signal to reduce the noise of a high-pressure-ratio axial flow compressor. The steps of LNF-MSVD are listed after determining the method's feasibility and effectiveness, as well as the amplitude and frequency standards of the excitation signal. Using this method, the acoustic mode was extracted accurately and clearly shown using an acoustic modal diagram, which was used to acquire experimental acoustic modal order. The superiority of LNF-MSVD for extracting acoustic mode was confirmed by a comparison with SVD and wavelet analyses. By comparing experimental acoustic modal order with theoretical acoustic modal order calculated through an empirical formula, the likely source of noise of the compressor was successfully located.