The microphone data collected in aeroacoustic wind tunnel test contains not only desired aeroacoustic signal but also background noise generated by the jet or the valve of the wind tunnel, so the desired aeroacoustic characteristics is difficult to be highlighted due to the low Signal-to-Noise Ratio (SNR). Classical cross spectral matrix removal can only reduce the microphone self-noise, but its effect is limited for jet noise. Therefore, an Airflow Background Noise Suppression method based on the Ensemble Empirical Mode Decomposition (ABNSEEMD) is proposed to eliminate the influence of background noise on aeroacoustic field reconstruction. The new method uses EEMD to adaptively separate the background noise in microphone data, which has good practicability for increasing SNR of aeroacoustic signal. A localization experiment was conducted by using two loudspeakers in wind tunnel with 80 m/s velocity. Results show that proposed method can filter out the background noise more effectively and improve the SNR of the loudspeakers signal compared with spectral subtraction and cepstrum methods. Moreover, the aeroacoustic field produced by a NACA EPPLER 862 STRUT airfoil model was also measured and reconstructed. Delay-and-sum beamforming maps of aeroacoustic source were displayed after the background noise was suppressed, which further demonstrates the proposed method’s advantage.

Acoustic source localization using distributed microphone array is a challenging task due to the influences of noise and reverberation. In this paper, acoustic source localization using kernel-based extreme learning machine in distributed microphone array is proposed. Specifically, the space of interest is divided into some labeled positions, and the candidate generalized cross correlation function in each node is treated as the feature mapped into the hidden nodes of extreme learning machine. During the training phase, by the implementation of kernel function, the output weights of the classifier are calculated and do not need to be tuned. After the kernel-based extreme learning machine (K-ELM) is well trained, the measured generalized cross correlation data are fed into the K-ELM classifier, and the output is the estimated acoustic source position. The proposed method needs less human intervention for both training and testing and it does not need to calibrate the node in advance. Simulation and real-world experimental results reveal that the proposed method has extremely fast training and testing speeds, and can obtain better localization performance than steered response power, K-nearest neighbor, and support vector machine methods.