APPLICATION OF PLANAR MICROPHONE ARRAYS AND COMPRESSIVE SENSING FOR LOCALIZATION OF COMPRESSED AIR LEAKAGES

摘要

Compressed air is a kind of expensive sources in modern industrial factories. Leakage is the largest waste of energy in a compressed air system. When air escapes through a small orifice of a compressed gas system, a turbulent noise spectrum is produced. A algorithm combined Orthogonal Matching Pursuit (OMP) algorithm and singular value decomposition is proposed for localization of compressed gas leakages. The method can show the locations of compressed gas leakages on super-resolution source maps in the low signal-to-noise (SNR) environment. The experiments are conducted in a mechanical laboratory, with the noise of an air compressor and environment noises served as the background noises. The SNR is very low in the laboratory. The leakage orifices are designed in the arbitrary positions in the three-dimensional space. The results obtained with the proposed method is compared with those with conventional beamformer (CBF) and the Tikhonov regularization (TIKR) method. The performances of the CBF method and TIKR method degrade due to low SNR in the laboratory. At the same time, the results show that the CS algorithm is computationally more effective and can present a super-resolution map. This work proves the feasibility of phased microphone array and a CS algorithm applied to the localization of compressed gas leakages.